Purpose: We introduced a multimodal training program for laparoscopic pyeloplasty (LP) and evaluated its safety, feasibility, and efficacy. Methods: The program consisted of box-trainer training, animal-model training, and operative training. Five trainees with different experiences in open pyeloplasty and laparoscopy were exposed to the program. The mentor performed objective and subjective evaluations at each stage to ensure the training quality and operation safety. The perioperative parameters of five groups of patients who underwent LP by the trainees independently were evaluated. Results: All trainees successfully finished the training program and independently performed five LPs under the supervision of the mentor. Five trainees spent different training times on the box-trainer and animal-model training,but acquired similar laparoscopic proficiency. There were no conversions to open procedures, transfusions, or deaths among the patients. No statistically significant difference was found in the operative time, estimated blood loss, postoperative hospital stay, and perioperative complications among the five trainees (P > 0.05). Conclusion: The multimodal training program can be used to train residents to perform advanced LP through step-by-step training from box trainer to animal model to clinical practice. The mentor-initiated approach is important to guarantee the training quality and safety.