In order to reduce the failure probability of drilling equipment in the process of operation, the application of Bayesian networks (BN) in fault diagnosis of drilling equipment system is proposed. A conditional probability table retrieval algorithm based on a genetic algorithm is adopted to improve the Bayesian network, random variables in the improved Bayesian network and connection relations among nodes in the network are described through the conditional probability table, a network structure and node parameters are defined, and an improved Bayesian network model is constructed; By calculating the system reliability and quantitatively analyzing the influence of the system equipment failure on the system operation reliability, a training and learning sample base containing historical failure types, a historical database and the operation parameters when the failure occurs is constructed, and the training and learning sample base is used as the input of the improved Bayesian network model to realize the fault diagnosis of the drilling equipment system. The experimental results show that the method can accurately diagnose the fault type of the drilling equipment system, and the fault diagnosis results can provide data support for the later maintenance of the system.