The current study describes the use of a procedure called the "potty party", an all-day toilet training method using basic learning principles of "errorless" discrimination with 3 children with incontinence diagnosed with autism. At the start of treatment days, each participant was greeted and prompted to request the bathroom. After the child requested the bathroom, they were taken to the restroom, pants and underpants were removed, the child was seated on the toilet. While on-seat, participants were given liquids, less-preferred reinforcers and engaged in typical daily activities such as discrete trial training. When the child voided in the toilet, specific verbal praise and highly preferred edible reinforcers were delivered. The child was also given time off the toilet with their most preferred toys or items/activities. Time off-toilet subsequently increased with each in-toilet urination until the participants were spending the same amount of time off-toilet, out-of-bathroom, and in the classroom as their peers. Results for all three participants using 3 non-concurrent A-B phase designs indicated that the potty party procedure was effective in decreasing accidents to zero levels for all three participants and increased in-toilet urination for two. Future directions for research in toileting are discussed.