Abstract: The realization of parallel computing for all kinds of P systems is the research hot point for membrane computing.Matrix representation for spiking neural P system with exhaustive use of rules is proposed in this paper.With the completion of the matrix representation, the authors research the GPU implementation of the spiking neural P system with exhaustive use of rules.The parallel computing simulation of the spiking neural P system with exhaustive use of rules gives the acceleration performance on GPUs.In ten times experiments, the average acceleration ratio of GPU to CPU is 1.4.