Aiming at the problem that the hydraulic system fault of mine column cannot be diagnosed quickly and accurately at present,a variety of fault diagnosis methods were proposed based on the optimization algorithm and establishing simulation model to analyze single fault mechanism.The column hydraulic system simulation model was established by combining the column physical module with the column hydraulic system module.The influence of single fault was analyzed based on Simulink,and the diagnostic model was established based on snake optimized LSTM (SO-LSTM) neural network.Finally,the model was verified by an example according to the actual data.The results show that using snake optimized LSTM neural network,the recognition rate of the hydraulic column fault simulation data is 99.5%,and the recognition rate of the real hydraulic column fault data is 97%,which is only 2.5% lower than the prediction accuracy of the model simulation data.The prediction accuracy is high,and the expected target is achieved.