Abstract: In order to solve the problem that the traditional stochastic resonance system can only process low-frequency signals and cannot directly process high-frequency seismoelectric logging signals, the dynamic equation of the nonlinear bistable stochastic resonance system is improved by using the phase trajectory time-scale transformation. By analyzing the numerical simulation model and frequency domain diagram of the numerical simulation model and the output signal of the circuit, it is found that the output signal of the system has two obvious stable states, The peak of the spectrum at the frequency of the signal under test is significantly increased, the amplitude and energy of the output signal are significantly enhanced, and the signal-to-noise ratio of the measured signal is improved. A stochastic resonance system with the output signal-to-noise ratio as the evaluation function and the genetic algorithm is used to optimize the system parameters to obtain the optimal output, and the output signal signal-to-noise ratio is improved by 23.5642 dB. Finally, the optimization system is applied to the seismic logging signal, and the amplitude at the eigenfrequency of the output signal of the random resonance system is 44 times that of the traditional linear filtering technology. The results show that The bistable stochastic resonance system based on the phase trajectory time-scale transformation can directly process the high-frequency seismic logging signal, weaken the noise in the signal, and significantly improve the clarity and quality of the signal.