Good path planning can greatly improve the working efficiency of industrial robots. To solve the problem of interference judgment between industrial robots and obstacles, an interference judgment method based on axial angle is proposed. Combined with interference judgment, the range of the obstacle point was found and expanded. Based on the improved path planning method of particle swarm fusion artificial potential field method, the gravitational field and repulsive force field models are constructed to generate the initial particle swarm. The traction operation is proposed to solve the problem that particle swarm optimization is easy to fall into local optimal. Bessel curve is used for smoothing operation to solve the problem that path points in grid environment are too many, which is not conducive to robot walking. An example shows that this method can overcome the shortcomings of traditional and intelligent algorithms in robot path planning, and it has strong searching ability and convergence ability.