The invention discloses an MEMS (micro-electromechanical system) gyroscope compound learning control method based on a parallel estimation model, and aims to solve the technical problem of poor practicability of an existing MEMS gyroscope mode control method. The technical scheme includes that neural network prediction errors are built according to the parallel estimation model and a dynamics model, compound adaptive laws of neural network weights are designed according to tracking errors, weight coefficients of a neural network are corrected, effective dynamic estimation of unknown dynamics is achieved, a system is indeterminate for parameters and insensitive to external interference in a sliding mode, and feed forward compensation of the unknown dynamics is achieved by the aid of a sliding mode controller. According to the method, the compound adaptive laws of neural network weights are designed, the weight coefficients of the neural network are corrected, effective dynamic estimation of the unknown dynamics is achieved, feed forward compensation of the MEMS gyroscope unknown dynamics is achieved according to sliding mode control theories, control accuracy of an MEMS gyroscope isfurther improved, and the method is good in practicability.