In order to reduce the observation time of X-ray pulsar signal identification, a new identification algorithm based on the photon distribution statistics is proposed. The influence of cumulation period variation on photons distribution is analyzed, according to whether the cumulation period is correct, two kinds of Poisson models are developed to characterize the cumulation profiles. The photon distribution statistics is constructed, and the difference of photon distribution statistics between two kinds of profiles is presented. Rossi X-ray timing explorer observation data is used for simulation, and the simulation result verifies the effectiveness of the suggested method.