The multi-signal flow graph model is the most widely used Testability model, but its ability to handle uncertain information is weak, and the failure information utilization rate is low, which cannot provide researchers with more data support. Moreover, the model cannot update the structure and parameters through learning. The model established may not be suitable for all stages of the entire life cycle of the equipment, and there is a risk of requiring multiple modeling. In order to solve the above problems, a Bayesian network was introduced and a MSFG-BN combined model was proposed. The multi-signal flow graph was used to reduce the difficulty of modeling the Bayesian network. The Bayesian network was used to increase the processing of uncertain information and improve the information utilization. At the same time, with the help of Bayesian network learning capabilities, the model can update its own structure and parameters, improving the applicability of the model at all stages. It has been verified that the model can play its role effectively.