NO_x emission of Diesel engine is the main harmful emission substance of motor vehicles; accurate measurement of NO_x emission is conducive to the control of urea injection to reduce emissions. However, the existing NO_x sensors and emission MAP obtained by calibration are both difficult to achieve real-time measurement of NO_x under transient conditions. Principal component analysis (PCA) was used to reduce the dimension of diesel engine operating parameters for world harmonized transient cycle (WHTC). A real-time diesel NO_x prediction model was built based on long and short-term memory (LSTM) neural network, and the parameters of LSTM were optimized by grey wolf optimization (GWO) algorithm. The results show that the mean absolute percentage error (MAPE) of GMOLSTM prediction model on the untrained data set is 3.23%, which proves that the model can accurately achieve real-time prediction of NO_x emissions of diesel engines. In addition, the model has good generalization ability and reliability, which provides a reference for the realization of diesel emission control with software instead of hardware.