In order to obtain the steady working point of compressor working state and build the relationship between compressor working state of near surge stall and compressor pressure signal, the near surge stall test for the low speed twostage axial compressor is carried out at different speeds under uniform inlet and distortion inlet to measure the pressure signals of the compressor from normal operation to stall condition in real time, and a compressor stall prediction algorithm based on autocorrelation coefficient is proposed in the time domain to predict and analyze the pressure signals of compressor inlet, outlet and the tip position of the head rotor. The results show that the proposed prediction algorithm has good stall prediction ability, and the blade tip, 20% from the blade leading edge, is most suitable as the observation points for the stall prediction of the compressor, where the compressor blade tip pressure self correlation data has obvious monotonic correlation with compressor surge margin.