With the advantage of large volume capacity, heavy goods vehicles have become an important part of road freight transport. However, the proportion of heavy goods vehicles involved in traffic accidents is very high, especially in traffic accidents with serious casualties. These traffic accidents have brought great harm to lives and property, so it is necessary to study the factors influencing the traffic accident severity of heavy goods vehicles and corresponding countermeasures. Based on traffic accident data of heavy goods vehicles in Shenzhen, the statistical analysis of heavy goods vehicle traffic accidents had been examined from the aspects of driver, vehicle, road, and environment. The characteristics of heavy goods vehicle traffic accidents have also been identified, and the influence factors of heavy goods vehicle traffic accident severity have been singled out. With the number of fatal accidents in traffic accident data as a reference, the main factors strongly related to the heavy goods vehicle traffic accident severity were determined through Grey correlation analysis. We then built the data set and the test set based on the selected main factors. We established a Bayesian network model to analyze the relationship between the heavy goods vehicle traffic accident severity and influence factors. The structure of Bayesian network was constructed. The parameter estimation of Bayesian network was conducted by EM algorithm, and the validity of the model has been verified. Then the reasoning analysis was carried out by applying this model. Based on the results of Bayesian network reasoning analysis, both the hidden dangers of heavy goods vehicle traffic accident and problems in the supervision and management of heavy goods vehicles were analyzed, and we proposed countermeasures of strengthening supervision and management and improving road hardware conditions to improve the traffic safety level of heavy goods vehicles.