Abstract: Bayesian kernel machine regression (BKMR) model is an advanced semi-parametric Bayesian method that has been widely applied in environmental epidemiology to assess the health effects of mixed exposure to multiple pollutants. Compared to traditional linear regression or weighted quantile sum regression models, the BKMR model can effectively capture nonlinear exposure-response relationships, detect interactions between pollutants, and address multicollinearity, making it a powerful tool for evaluating the joint effect of exposure to various pollutants. This study provides a comprehensive overview of the theoretical foundation of BKMR model and its implementation procedure in R. Additionally, this study demonstrates the application of BKMR through multiple practical case studies, highlighting its utility in analyzing the effects of mixed exposure to heavy metals on liver, thyroid, and renal function. Through step-by-step code demonstration and graphical interpretation, we provide a detailed explanation of the model output and its public health implications. Although BKMR demonstrates considerable potential in health effect assessments, it still faces challenges such as high computational demands, steep learning curve in result interpretation, and limited visualization clarity. Therefore, we also discuss recent methodological advancements such as knot compression and algorithm optimization. In conclusion, BKMR model provides robust support for evaluating health risks associated with mixed exposure to multiple pollutants and provides a theoretical and technical foundation for methodological innovations and policy formulation in environmental health.