Abstract: In order to explore the influence of process parameters on the grinding runout of the end face of the turntable for sub-micron grinding. The improved sparrow search algorithm-back propagation (ISSA-BP) was proposed to establish a runout prediction model for turntable end grinding. The ISSA-BP neural network model was constructed with grinding wheel speed, grinding wheel feed depth, workpiece feed rate, grinding cycle times, and end face runout value before grinding as the input layer, and post-grinding runout value as the output layer. The grinding experimental data was substituted into the model for progressive comparison to verify the effectiveness of the improved algorithm, and the prediction superiority of the model was verified by comparison with the genetic algorithm GA-BP (genetic algorithm-back propagation) neural network. The experimental comparison results show that the prediction accuracy of the improved prediction model is 92.11%, which can accurately predict the grinding runout of the end face of the turntable. The prediction model was used to predict each single factor, the influence law of each parameter was analyzed, and the experiment was designed for comparison and verification. The turntable end grinding runout prediction model can realize the five-factor post-grinding runout value prediction for end grinding, and the single-factor post-grinding runout value can be predicted for each parameter, and the influence law of submicron-level micro-grinding process parameters on end-face runout of the turntable end face is obtained.