As one of the important agricultural equipment, the design of harvesting machinery has many typical characteristics, such as complex and diverse structures, strong seasonality, and obvious characteristics of users diversity and customization requirements. The traditional R&D mode have some problems, such as long design cycle, low efficiency and difficult quality assurance. Taking the ear peeling device of corn combine harvester as the research object, according to the relationship among the structural characteristics, technical parameters and performance evaluation indexes of the peeling device, the design method of the ear peeling device based on knowledge was presented. Firstly, the design process of peeling device was clarified, and the modular design scheme was formulated, which was divided into special parts module, general parts module, and standard parts module according to functions. The special module was the core component of the peeling device, mainly including the peeling roller and the pressure feeder; the general module included the feeding roller, the conveying mechanism, the impurity eliminator, the driving mechanism and the fruit spike recovery mechanism and so on; the standard module included the driving part, the connecting fastener and the bearing and so on. Secondly, according to the scope of standards, rules and constraints, the related design knowledge base of peeling device was established, the mathematical relationship between the features, operation form corn varieties, transmission scheme, structure parameters, and operating parameters were analyzed. The representation and storage method of the design knowledge of corn ear peeling device was studied. Meanwhile, the framework representation method was used to decompose the husking plant, and a topdown hierarchical structure was designed. Finally, the computer aided design (CAD) and knowledge engineering were integrated to study the reasoning mechanism design of corn ear peeling device. In addition, the working performance evaluation model of corn ear peeling device was established, including the component life calculation model, the bract stripping rate evaluation model, the grain loss rate evaluation model, and the grain breakage rate evaluation model. Based on the Visual Studio platform, the knowledge base, reasoning machine, evaluation model and humanmachine interface of system were integrated, and a design system of corn ear peeling device with knowledge was developed, which realized the realtime calculation output and parameter evaluation of design parameters under the input of users demand parameters. On the basis of the above research, taking the parameters of TPJ16 corn ear peeling device as an example, the input power at the interface was 7.5kW, the feeding amount was 16.6t, the key structural parameters and motion parameters of the peeling device were calculated and evaluated, and the performance of the design parameters was evaluated. The solution results showed that the peeling rate of bract leaves, grain crushing rate and grain loss rate of the peeling device were 96.01%, 1.42% and 3.25%, respectively. The research results provided a reference for the rapid design of agricultural machinery and equipment.