Abstract: In this paper, the electromagnetic detection platform based on magnetic Barkhausen noise(MBN)principle is built, and the stress detection is carried out with Q235 steel as the experimental object. Before and after the test, different waveforms, frequencies and voltages were used to test the relationship between the peak value of MBN signal and frequency and voltage, and determine the optimal excitation parameters. The relationship between the MBN signal and the tensile stress is determined by the static load tensile test. The relationship between the peak value of the MBN signals and the tensile stress is fitted and the feasibility of using the peak value of the signal as the stress evaluation index is analyzed. The results show that the MBN signal generated by the sine wave as the excitation signal has obvious characteristics and the SNR is high; the peak value of the MBN signal increases first and then decreases with increasing tensile stress; the nonlinear fitting curve can be used to evaluate the stress concentration.