As a preprocessing step of target segmentation, superpixel can greatly reduce the amount of subsequent data processing, and plays a vital role in image segmentation. In most superpixel algorithms, seed points are sampled on a regular grid or initialized randomly, which easily leads to undersegmentation. In order to obtain a good distribution of seed point and avoid undersegmentation, an adaptively initializing superpixel seeds method based on Kmeans++ is proposed and used to improve the algorithms of SNIC. The experimental results show that the improved SNIC algorithm can get higher boundary recall rate and lower undersegmentation error rate than that of the traditional algorithm without a lot of computational cost.