Abstract: To accurately and quickly identify the slope value of the road with steep slope and large slope change rates, a slope recognition algorithm was proposed based on the multi-information data fusion filtering. The advantages and disadvantages of different slope recognition algorithms were analyzed. The slope recognition models based on the vehicle dynamic, the acceleration sensing information with considering slope change rate and the GPS were respectively established. The interactive multi-model Kalman filter algorithm（IMM-KF） was adopted, and the three slope recognition models were jointly filtered and estimated. The participation ratio of the slope recognition model was adaptively adjusted under different operating conditions. Taking the multi-sensor information of in-wheel motor vehicle as carrier, the dSPACE test platform covering the dynamic model of four-wheel independent drive electric vehicles was constructed, and the simulation was completed. The results show that under the conditions of constant slope change rate, continuous slope change and standing slope, the slope identification results can quickly and accurately follow the actual value after small oscillation, which indicates that the proposed algorithm can improve the accuracy and robustness of slope recognition.