Sowing depth is one of the important indicators for evaluating the quality of sowing, which is directly related to the seed germination and plant emergence. In order to achieve the optimum yield of agricultural crops, it is necessary to monitor the sowing depth in real time. Because of the inappropriate response of the machine dynamics to harsh soil conditions, such as the compacted soil undulations and the presence of the stubble, the existing measurement methods of sowing depth cannot meet the requirement of measuring accuracy, and the lack of an effective realtime sowing depth evaluation system affects the improvement of realtime operation quality. Moreover, the development of intelligent seeder puts forward higher demand for existing communication mode. To solve these problems, a sowing depth monitoring and evaluation system based on CANbus was proposed. The system was mainly composed of a tablet personal computer for realtime monitoring, an electronic control unit (ECU) for data acquisition and a sowing depth measuring device. The sixrow seeder was taken as the research object. On the basis of expounding the working principle of sowing depth measurement, the sowing depth measuring device based on the swing angle of the gauge wheel was designed and the corresponding measuring models were established. In addition, an intelligent bus communication protocol based on ISO 11783 standard was made, and a realtime monitoring and evaluation upper computer interface based on LabVIEW was developed. In order to study the variation characteristics of sowing depth, twofactor split plot experiment was carried out with tillage mode and working speed as experimental factors. Spectrum analysis of the data showed that the main frequency amplitude of the change of sowing depth was decreased with the increase of working speed, and the variation range of sowing depth under notillage was larger than that under rotary tillage, while the oscillation frequency was less affected by the change of tillage mode and working speed, which was mainly concentrated below 0.4Hz could provide a reference for lowpass filtering of signals in subsequent sowing depth control. Further, a field test was conducted to test the system performance. The field test results of monitoring and evaluation showed that the obtained monitoring map of sowing depth could characterize the soil information of the plot, such as compactness, which was favorable for subsequent seeding decisionmaking. Compared with the manual measurement method, the stability and consistency of the system monitoring were better when the speed was 6~10km/h. Within the industry standard error range of 10mm, the maximum values of average sowing depth, qualified rate, standard deviation (SD) and coefficient of variation (CV)were 50.01mm, 78.95%, 8.95mm and 17.90%, respectively, the relative errors of average sowing depth, qualified rate, SD and CV were 4.20%~9.74%, 6.11%~17.92%, 10.93%~16.32% and 18.83%~19.79%, respectively, which met the needs of monitoring and evaluating of the actual sowing depth.