How to Stop a Big Dog from Jumping on People

A 40 kg (88 lb) dog jumping on people is a physical incident, not a behavior quirk to laugh off. The reset that works: no eye contact, no touch, no words the moment those paws leave the floor, plus a firm sit before greetings resume. Add stranger-management at the front door and a leash-tether backup for the guests who forget the rules. Six weeks of that beats every spray-bottle trick on the shelf. Knee-in-chest is out. Management is in.

A shepherd, a rottweiler, a mastiff at full excitement can put a full-grown adult on the floor. The dog is not being bad; the dog is being 40 kg (88 lb) of muscle without brakes. Ten minutes a day, six weeks of household consistency, and the front door stops being a rugby ruck. What follows is the whole protocol, from the first yard rep to the guest at the door.

What you'll need

• A flat lead, four to six feet (1.2 to 1.8 m). Never retractable for a jumper.
• A properly fitted harness or a flat collar the dog cannot back out of.
• Small high-value treats, the size the dog swallows in one chew.
• A tether point at the front door: a wall hook, a heavy anchor, or a d-ring screwed into a nearby stud.
• One household that agrees on the rules, and one line to say to guests at the door.

The gear list is short on purpose. If the walk itself is a fight, the harness sizing question is worth reading first: a barrel-chested dog needs different geometry than a Labrador-shaped one, and our loose leash walking guide covers the sit-and-greet rehearsal work that this protocol builds on.

Step 1: Fix the reward first

Before technique, motivation. Dogs jump because it works. Every 'off, off, off' while a hand pushes them down is attention. Every look in the eye is a win. Even the shove is a game to a dog who wants contact. Jumping is a slot machine, and every person in the house has been feeding it quarters without realizing.

The whole protocol below only works once the household agrees on this: from today, jumping earns exactly nothing. Not a word, not a touch, not a glance. That is the reward change the rest of the training depends on.

Step 2: Teach sit-before-greeting in the yard

The trade is straightforward. Jumping used to earn attention. Now sitting earns it. Ask for a sit as you approach the dog. The moment their bottom touches the floor, the treat and the calm greeting land. The moment the dog stands up before released, the greeting stops and resets.

Run this a hundred times in the yard before the front door is anywhere in the plan. Muscle memory built in calm becomes possible under stress. Common mistake at this stage: rewarding the sit that lasts a millisecond. Wait two full seconds of parked-butt before the treat lands. Short-hold sits collapse the moment a real guest walks in.

Step 3: Do the turn-and-ignore reset

The workhorse move. The moment paws leave the floor, become a boring wall. No eyes, no words, no touch. Rotate ninety degrees away so the dog sees your back. Wait for four paws on the ground for two full seconds, then re-engage. If the dog jumps again on re-engagement, turn again. Same reset, every time.

Why turning away works: dogs are tuned to what earns attention and what loses it. Turning is not read as punishment; it is read as consequence. Same result, cleaner mechanism. Common mistake at this step: sneaking a glance to see if the reset is working. Dogs read the glance. Fully turn, fully wait.

Step 4: Set up the front door

The front door is where a hundred yard-reps go to die. Set the door up so the dog cannot lose. Two options that both work.

Option one is the tether. Clip the lead to a wall hook or a heavy anchor six feet (1.8 m) from the entry. The guest arrives, the dog is tethered, and the guest ignores the dog until sit-and-hold lands, then approaches to greet. Option two is the mat cue. The doorbell rings, the dog goes to a mat or crate on command, the guest enters, and the dog is released to greet only after a settled hold. Whichever way, the dog cannot ambush the door.

Common mistake at this step: skipping the setup because 'the friends love the dog.' The friends love the dog until the day the dog knocks a grandmother down and someone breaks a hip. Set the door up for the worst-case guest, not the best-case one.

Step 5: Recruit the household and the guests

Big-dog jumping is a household training problem, not a dog training problem. Everyone in the house is either running the protocol or feeding the slot machine, and a single family member undoing the work is enough to reset the whole clock.

Guests need coaching at the door, not after the jump. The line is polite and firm: 'we're working on this, so if she jumps, please turn away.' Say it before the guest walks in. Say it again on the next visit. Some guests will not listen; those visits get the tether treatment until the household is confident the dog can meet an untrained guest without launching.

Troubleshooting

The reset works at home but fails at the door. Rebuild at the door specifically. Set up a fake-guest rehearsal three times a week with a family member playing the guest, ringing the bell and walking in. The dog needs practice at the exact moment the trigger fires, not a version of it.

The dog jumps as soon as the human turns. Add a full step forward as the turn happens, breaking the launch geometry. If the dog still connects, the tether goes on at the door for the next two weeks with no exceptions.

The dog does not jump on the family but jumps on visitors. Common with dogs who read family as calm-time and visitors as party-time. Same protocol, run only during visits. Every visit is a training session for the six-week window.

Nothing has moved after two weeks. Two weeks is not the timeline. Six weeks is. If nothing has moved after six weeks of daily practice with all household members consistent, two sessions with a certified positive-reinforcement trainer is money well spent. Not an aversive trainer, a positive one.

The knee-in-chest question

Skipping the older knee-to-chest correction bothers some readers, so it is worth naming. That correction is out for three reasons. It hurts the dog for a habit the human taught. It confuses the mechanism, associating pain with the human rather than with the jump. And it occasionally goes wrong at the wrong angle and injures the dog. The turn-and-ignore reset works better, does not risk the dog, and does not risk the friendship.

What good looks like at six weeks

At six weeks with a household that stayed consistent, the dog approaches guests, sits, gets greeted, and moves on. Not perfect every time. Better every week. On a big-excitement day (a favorite person visiting for the first time in months), one jump may still happen. Reset, sit, greet. The point is not zero jumps forever; the point is a protocol the dog knows, so the failures become the exception instead of the rule.

Next steps

If the front door is one problem, the walk is often the other. Same physics, same principle: change what the dog is being paid for. Our two-second rule for pullers picks up where this leaves off.

For the wider training set (the sit, the leave-it, the recall trio, plus the reactivity work that lives alongside the door protocol) the training pillar collects the pieces in the order to read them.