Particle Swarm Optimization is a new efficient and parallel optimization algorithm. Because of its intelligence，simpleness and easy operation，this means has aroused much attention among scientists. Though PSO has been applied to many fields, its application in geophysical inversion is very insufficient. The joint inversion of surface and borehole magnetic data is a new geophysical inversion which combines the advantages of surface and borehole magnetic data. This paper created a vertical cube model and utilized PSO to the theoretical model test and the joint inversion of surface and borehole magnetic data in the Daye exhausted mine, and the result shows the feasibility of PSO in geophysical inversion and the superiority of the joint inversion of surface and borehole data.