The development of 5G network has promoted the strategic transformation and upgrading of the mining industry, which can reduce the safety risks of the mining industry and improve the efficiency of mining production. The intelligence mine 5G SA network sinks the user plane function to the mining area, and it leaves the control plane in the center of the large area. All data services such as remote control, 5G video backhaul, VR/AR inspection, and data collection are accessed through the 5G edge computing interface to the mining area application servers. This paper proposes the PSO_I algorithm, and introduces the Bayes evaluation method to calculate the average particle isolation factor, which is used as a constraint to improve the shortcomings of traditional PSO. The isolation fitness function is constructed based on the network slicing orchestration overhead and the utilization of physical node resources and link resources, and the slicing orchestration scheme is quickly found by iterative solution. After the actual 5G network slice arrangement test in the mining area, the results show that the PSO_I algorithm can improve the slice acceptance rate, the benefit-cost ratio and the link resource utilization rate, and improve the network uplink rate while reducing the execution time. Compared with the SecPSO, Slice and SSPS algorithms, the network uplink average rate optimized by the PSO_I algorithm is increased by 19%, 48% and 52%.