The invention belongs to the technical field of valve fault detection and diagnosis, and particularly relates to a fault diagnosis method and diagnosis system for an electric valve actuating mechanism. The method comprises the following steps of acquiring input and output variables of a BP neural network model by making utocorrelation analysis on the electric valve actuator; training the model toobtain an initialization value of the model; optimizing the model by using a genetic algorithm to obtain an optimal prediction value of the model; and collecting true value of the valve position opening degree of the electric valve actuating mechanism, and finally adopting a residual square between the true value and the predicted value as a fault state index of an electric valve actuating mechanism. According to a working principle of an internal structure of an electric value actuator, the working state is analyzed. The BP network is optimized by utilizing global search optimal characteristic of the genetic algorithm. The defect that the BP algorithm falls into local optimum in learning is avoided. The network has good convergence and adaptability, and the network has a good fault recognition effect.