According to the difficulty of spectral image clustering, an image clustering algorithm is proposed based on fuzzy kernel P system. Through the analysis of fuzzy clustering and kernel clustering, the fuzzy kernel system is built to optimize the organizational P system, and all cells are coordinated to calculate the optimal clustering center in the mapping space to achieve parallel computing. It can reduce the sensitivity of the initial clustering center, improve the global search ability of the algorithm, avoid falling into the local minimum, and improve the clustering performance of the algorithm.