The faults in CNC machining can affect machining accuracy.A prediction method was proposed to predict the faults of the machine tool in advance before processing and avoid the impact of machine tool faults on processing accuracy.In order to quickly and accurately predict the faults of CNC machine tools,genetic algorithm was used to optimize the long short term memory neural network model to predict the faults that might occur in CNC machine tools in service.The fault signals under different states were collected as samples for network training,and the network model was used to predict the state of machine tool fault.The results indicate that GA-LSTM is a high-precision prediction model with good performance in CNC machine tool fault prediction,the machine tool faults that affect the machining accuracy can be avoided.