Aiming at the problem of not directly extracting the fault from multiple frequency couplings in the actual working conditions of the planetary gearbox,a fault diagnosis method based on salp swarm algorithm optimized variational mode decomposition (SSA-VMD) combined with 2.5D spectrum was proposed.The SSA was used to optimize the parameters of VMD; the decomposed signal was reconstructed to reduce noise interference by the autocorrelation coefficient; and finally the 2.5D spectrum was used to decouple the frequency coupling in the reconstructed signal.A planetary gearbox wear fault full life cycle test bench was set up to collect vibration signals,the proposed method was used to decouple the frequency components of the faults participating in the coupling,and the evolutionary law of the wear fault of the planetary gearbox was revealed.The research results show that with the deepening of the degree of wear faults,the frequency of wear faults characteristic increases significantly.