The invention discloses a component behavior model mining method and device. The method comprises the steps: S1), operating software comprising components, dynamically collecting a parameter behaviorinteraction sequence of the components, and forming a sequence set; S2) combining the same component behavior interaction sequences with different parameter values; S3) constructing a tree based on the combined component behavior interaction sequence; S4S4) merging the equivalent nodes in the tree to obtain a finite-state machine R '; S5) concluding invariants of the parameters according to the parameter observation value set to serve as daemon conditions of corresponding edges in the finite-state machine R '; S6) calculating the probability that the component behavior in the finite-state machine R'meets a parameter invariant; and S7) obtaining a final component behavior model represented by the parametric probabilistic automaton based on the probability of migration in the step S6). According to the method, the dependency relationship between parameters and component behaviors in the component behavior model is considered, and the noise in the model mining process is effectively processed by adopting the probability model, so that a more accurate component behavior model can be obtained.