Due to the fixed vision and step of artificial fish swarm algorithm resulting in algorithm optimization speed slow, easy to fall into local optimum value, it introduces a variable coefficient factor for adapting the vision and step in artificial fish’s swarm, rear end and foraging behavior; in addition, to reduce computational complexity of late algorithm and obtain more effective artificial fish, adding a maximum number of iterations elimination mechanism. Then, using improved artificial fish swarm algorithm to optimize the kernel function and penalty coefficient of support vector machine, and applied it in short-term wind power prediction. The simulation results show that the improved artificial fish swarm algorithm optimization support vector machine has a better effect on short-term wind power prediction.