Objective To evaluate the predictive validity of Braden Q pediatric pressure ulcer risk assessment scale (Braden-Q scale) in severe pediatric patients. Methods The prospective cohort study design was adopted, and 111 samples were collected from intensive care units of three pediatric hospitals in Shanghai with convenience sampling. Two nurses were appointed as the data collector, and were in charge of the scoring of Braden-Q scale and skin assessment. Results Seven patients (6.3%) developed pressure ulcers. The cutoff score of Braden-Q scale was 17 for the risk of pressure ulcer, with the sensitivity of 0.571, specificity of 0.365 and area under receiver operating characteristic curve of 0.547. Conclusion The predictive validity of Braden-Q scale in severe pediatric patients is ordinary, and there is room for improvement.