The work aims to investigate the effects of power frequency, current intensity and steel plate moving speed on temperature field of the steel plate during induction hardening by numerical simulation, to provide a reference for the selection of parameters in practical application. The finite element calculation model of continual induction hardening process for steel plate was established by ANSYS APDL language to simulate the temperature field of steel plate under different process parameters numerically. The induction hardening experiment was carried out to 20 mm thick 40Cr steel plate with optimized process parameters. The temperature of key point on the steel plate was measured by a thermocouple. The microstructure and hardness of quenched steel plate were analyzed by metallographic microscope and microhardness tester. The maximum error rate between the calculation results and the measurement results of key point temperature was about 4%, which indicated that the model had high calculation accuracy. The analysis results of the temperature field under different process parameters showed that: the higher the power frequency and current density were, the faster the heating speed was. With the increase of power frequency, the depth of high temperature region increased at first and then decreased. As the current density increased and the moving speed of steel plate decreased, the depth of high temperature region became large. The microstructure in thickness direction of steel plate after quenching was basically divided into three regions: phase transformation hardening region, heat affected transitional region and untransformed region. The microstructure of phase transformation hardening region was fine acicular martensite with the highest hardness of 700HV, and the depth of hardened layer was about 6 mm. The martensite in heat affected transitional region decreased gradually, while untransformed region was still original pearlite and ferrite. All the simulation results coincide with experimental results well, which can be used to guide the selection of parameters in practical application.