Abstract: Abstract: Moisture content of grain is one of the greatest significant factors affecting the quality of food, which still plays an important role as an indicator in food storage, acquisition, processing,transportation, and so on. Recently, as the rapid development of the intelligent combine harvester, intelligent harvest and yield measurement have become possible. It is necessary to monitor the moisture content of grain in real time in automatic harvest. In addition, only when the weight of the harvested grain is converted to the weight of the grain with a fixed moisture content, the yield information can be measured more accurately, and the cause of loss rate can be analyzed more comprehensively. Comparing the domestic machine with that in other countries, the fact is that some latest combine harvesters designed in other nations have installed the real-time grain moisture content monitoring system, while the majority of the domestic combine harvesters have not yet used this technology in practice. By virtue of the huge cost of introducing the measurement systems from other nations, the import of this machine has not been applied in agriculture industry universally. Therefore, in the light of this situation, the research on the on-line monitoring of the grain moisture content for the combine harvester will be of great value in the domestic area to some extent. In order to analyze the influence of edge effect on the capacitive plate, a finite element analysis software COMSOL Multiphysics was adopted to establish the capacitive plate and its working environment model. The influences of the thickness of the capacitive plate, the plate spacing and the relative area on the edge effect have been evaluated. The three-factor and three-levelorthogonal experiment optimization simulation was carried out, using capacitance ratio and sensitivity as the evaluation indicator. According to the simulation results, a copper plate with a plate thickness of 0.15 mm, a plate spacing of 20 mm and a relative area of 3 000 mm2 between the plates was selected. Based on the STM32F103 series microprocessor chip, an on-line monitoring system of grain moisture has been constructed. This system included the power supply module, high frequency excitation signal module, AC (alternating current) small signal amplification circuit, capacitor plate, signal conditioning circuit, mean response power detector, and other designed sensor detection circuits. For the aim of the more accurate consequence of the monitor with respect to the moisture content of cereals, simplifying the circuit structure and reducing the cost, Multisim simulation and experimental verification were carried out on various frequency sources respectively. Then, the high frequency signal with the frequency of 10 MHz was selected as the excitation signal of the monitoring device. Finally, in order to get the changing relationship between the moisture content and the voltage, the output signal of the circuit was converted from a non-sinusoidal wave caused by the capacitor charge and discharge to a DC (direct current) voltage by a mean response power detector taking the AD8361 as a core. Moreover, the device could monitor the grain moisture content on-line, and real-time display and store information The indoor static monitoring and on-field monitoring test were conducted withthe on-line monitoring device of grain moisture content. As a result, the maximum relative error of indoor static monitoring test was 1.57% and the maximum relative error of field monitoring test was 2.07%.