Abstract Background and Aims Previous studies mainly focused on linear association between body mass index (BMI) and the risk of chronic kidney disease (CKD). In the current research, we conducted a prospective cohort observational study and Mendelian Randomization analysis to explore whether non-linear relationship exists between BMI and CKD. Method In the prospective cohort study, 495,465 participants from the UK biobank were included. The Cox proportional hazards model with fractional polynomial transformation was used to estimate the hazard ratio of BMI for incident CKD. In addition, nonlinear Mendelian Randomization analysis was performed using the fractional polynomial method to verify the causality of the association. Results In the observational analysis, a J-shaped association was observed between BMI and the risk of CKD. The risk of CKD was lowest when the BMI was around 18-20 kg/m2, and the hazard of developing CKD increased in participants with lower or higher BMI (Fig. 1). When assessing the causal relationship, in Mendelian Randomization analysis, despite that a nonlinear relationship still held true (P-value for non-linearity = 0.008), the odds ratio for CKD monotonically rised with increasing BMI (Fig. 2). Conclusion The J-shape association between BMI and CKD revealed by the observational cohort study was not supported by the Mendelian Randomization, which indicated a monotonically causal relationship that the risk of CKD rised with increasing BMI.