Abstract: To mitigate time delay and high resource consumption of edge computing in discrete manufacturing data processing, a task scheduling method for edge computing based on an improved gray wolf algorithm is proposed. By improving the nonlinear convergence factor and the dynamic weight, the optimization speed and accuracy of the grey wolf algorithm are improved, and the terminal and edge terminal resource loss and the delay of task processing are effectively reduced. Experiments on processing delay and resource consumption under different data tasks prove the effectiveness of the proposed model. Compared with the three main task scheduling algorithms, the proposed model achieves the lowest data processing resource consumption and delay. Furthermore, by combining edge computing task scheduling with an intelligent optimization algorithm and applying it to discrete manufacturing, the task processing speed of equipment is improved, and the energy consumption is reduced. This model serves as a reference for the intelligent transformation of discrete manufacturing.