Abstract: In order to study the contact parameters between wheat and conveyor belt used in belt conveyor, a high-speed camera was used to capture collision process, and according to the motion geometry principle, the motion posture of wheat after the collision with the belt was described. Based on the high frame number of the camera, the actual initial rebound velocity of wheat was obtained, and then the coefficient of restitution of wheat-belt was acquired. By placing 100 wheat grains in a fixed position on the inclination belt, the actual average rolling distance of wheat was obtained with image processing. Different coefficient of rolling frictions were input into EDEM to obtain different average rolling distances in the simulation process, and compared with the actual situation to obtain the coefficient of rolling friction of the wheat and belt. A total of four sets of inclination angles (30°-60°) were set in the experiment. The coefficient of static friction was determined by the slip experiment of wheat and the same inverse reasoning method. The inclination angle of the belt when 30 grains with a uniform size slipped on the inclined belt in the actual situation was recorded. In the EDEM environment, the simulation experiment was carried out, and the inverse tangent function fitting was carried out in the Origin software. The actual inclination angle of the belt when the wheat slipped was substituted into the function and the actual maximum static friction coefficient of the wheat belt were obtained. The Box-Behnken experiment was designed to verify the reliability of contact parameters based on wheat stacking angle as the response value. The actual measured value of the stacking angle was set as the target, the optimal parameter combination was established and compared with the 0 horizontal combination. Through the above experiment, the results showed that the coefficient of restitution decreased gradually with the increase of the falling height in the range of 300-600 mm, and the results obtained at different collision positions were different. Combined with the actual working conditions of the belt conveyor, the expected value of the wheat-belt coefficient of restitution was 0.418 5. The coefficient of rolling friction was around 0.137 1, the relative error of the experiment was less than 5.00%, and the rolling friction coefficient obtained at the four inclination angles was the same, indicating that the coefficient of rolling friction was independent of the inclination angle. The maximum coefficient of static friction was around 0.503 0, the relative error was less than 2.04%. The influence factor values of the optimal solution were all about the middle level (0) through orthogonal experiment. The optimal parameters combination was input to EDEM for simulation experiment, the relative error of the stacking angle was 1.10%, suggesting that the three contact parameters of wheat belt were reliable. Experimental calibration in the present study provided a basis for the simulation study of belt conveyor for tilting wheat conveying and/or increasing wheat yield.