Aiming at the problem that working condition is complex in fact so that it is difficult to extract the gear fault feature frequency,a method of gear fault diagnosis based on variational mode decomposition（ VMD） and multi- feature fusion is proposed. Firstly,the mechanical vibration signal is decomposed into a series of modes by using VMD method. Then,the permutation entropy（ PE） value and energy of the first four modes of high frequency is calculated. Finally,the high- dimensional feature vectors composed of PE and energy are transformed as input of least squares support vector machine（ LS- SVM） to identify gear fault types.The experimental results indicate that VMD can better separate complex multi- components signal from each other,the PE and energy feature can reveal the gear fault information from different scales,the intelligent fault diagnosis method based on VMD and multi- feature fusion can identify the operating conditions of gear accurately and provide reference for the gear fault warning and severity.