Abstract: Abstract: A header profiling system can be used to stabilize the header height during the operation of a combine harvester. The harvesting efficiency can be improved to protect the header from damage by stones and other debris in fields. It is very necessary to enhance the control accuracy of header height via the design of the mechanical structure, rather than only the header profiling algorithm so far. Therefore, this study aims to improve the sensitivity of the header profiling mechanism response to the topography change in the field, and the accuracy of field terrain profiling. A main-sub-plate pressing profiling mechanism was proposed, which was composed of a main profiling plate, an auxiliary profiling plate, a spring, an angle sensor, and a four-link transmission mechanism. Firstly, the geometric and mechanical models were established to determine the key parameters. Next, a quadratic regression universal rotary combination test was designed to optimize the key design parameters of the profiling mechanism. Among them, the test indexes were taken as the supporting force of header to the auxiliary profiling plate, the supporting force of soil to the main profiling plate, and the soil subsidence distance, whereas, the test factors were the spring stiffness coefficient, the initial length of spring, the weight of auxiliary and main profiling plate. A particle swarm optimization was selected to determine the level range of each factor. A multi-body dynamics and discrete element method (MBD-DEM) coupling was utilized to implement the optimization test for the MBD-DEM models of the main-subplate pressing profiling mechanism. A mathematical regression model was then established between each factor and index. It was found that the stiffness coefficient and the initial length of spring were the main influencing factors in the three response indexes. At the same time, a multi-objective comprehensive optimization was performed on various factors. The optimization goal was to make the supporting force of the header to auxiliary profiling plate as large as possible, whereas, the supporting force of soil to main profiling plate and the soil subsidence distance as small as possible. An optimal parameter combination was achieved as follows: the spring stiffness coefficient was 464 N/m, the initial length of the spring was 0.09 m, the length values of auxiliary and main profiling plate were 484, and 450 mm, respectively. The width values of the main and auxiliary profiling plates were 200 mm, and the material was 301 stainless steel. The spring material was carbon spring steel wire, where the wire and outer diameters were 2, and 20 mm, respectively, and the number of effective coils was 45. The field experiment results showed that the supporting forces of the header to the auxiliary profiling plate were all greater than 0 in the three repeated tests, indicating that the profiling mechanism was always in a normal working state. The average values for the supporting force of soil to the main profiling plate were 85.23, 85.80, and 86.08 N, respectively, which were basically consistent with the predicted values of the regression model, indicating that the profiling mechanism was sensitive to terrain changes. The average values of soil subsidence distance were 5.4, 7.1, and 6.4 mm, respectively, which were less than 10 mm, indicating the high accuracy of field terrain profiling. Consequently, the field experimental data was better consistent with the predicted values of the optimization model, indicating the reasonable parameter design of the profiling mechanism. Anyway, the reliable main-subplate pressing profiling mechanism can fully meet the requirements of soybean harvesting. The finding can also provide a strong reference for the profiling system of header in the combine harvester.