In order to increase the detection speed and to improve the detection rate in low signal-to-noise ratio, a novel detection algorithm for pulsar integrated pulse profile is proposed based on S-transform. We prove that the local S-transform spectrum of Gaussian white noise follows chi-square distribution with two degrees of freedom in this paper, and based on this, threshold processing is operated on the S-transform power spectra of cumulative signal; then, the total energy of the time-frequency power spectral after threshold processing is accumulated to construct the detecting statistics to obtain the statistic decision. Moreover, the time-frequency power spectra after threshold processing can also be used to measure the time delay of pulsar signal. The simulation result verifies the effectiveness of the proposed method which is superior to a similar method based on the Gaussian distribution; in addition, it can give the time delay with a certain accuracy.