In the food production, operating environment, personnel, equipment and many other factors may introduce microbiological hazards into food, which makes consumers face health and safety risk. Effective implementation of risk management is essential to ensure food safety. In order to search, assess, prevent and control of varieties of risk factors, modular process risk modeling method is improved by adding hazard transfer process and control process. In the improved method, risk factors are abstracted as hazard transfer process. Control process characterizes the control measure on risk factors. Bayesian belief network is used as the model structure. Combined with predictive microbiology, the number and emergence probability of a microbial hazard in each process of food production is estimated using Bayesian inference. Simulation results show that the improved method can assess microbial risk in processing, and also may assess the impact degree of risk factors on the food product safety, trace the origin of microbial hazard, and help to choose preferred control measures.