Guideline compliant CPR is associated with improved survival for patients with cardiac arrest. Conventional Basic Life Support (BLS) training results in suboptimal CPR competency and skill retention. We aimed to compare the effectiveness of distributed CPR training with real-time feedback to conventional BLS training for CPR skills in pediatric healthcare providers. Healthcare providers were randomized into receiving annual BLS training (control) or distributed training with real-time feedback (intervention). The intervention group was asked to practice CPR for 2 min on mannequins while receiving real-time CPR feedback, at least once per month. Control group participants were not asked to practice CPR during the study period. Excellent CPR was defined as 90% guideline-compliance for depth, rate and recoil of chest compressions. CPR performance of participants was assessed (on infant and adult-sized mannequins) every 3 months for a duration of 12 months. CPR performance was compared between the 2 groups. A total of 87 healthcare providers were included in the analyses (control n = 41, intervention n = 46). Baseline assessment showed no significant difference in CPR performance across the 2 groups. The intervention group has a significantly greater proportion of participants with excellent CPR compared with the control group on an adult sized mannequin (14.6% vs. 54.3%, p < 0.001) and infant-sized mannequin (19.5% vs. 71.7%, p < 0.001) at the end of the study. In the intervention group, all CPR metrics except infant depth were improved and retained over the course of the study. Distributed CPR training with real-time feedback improves the compliance of AHA guidelines of quality of CPR.