Automatic mechanical tracking of rice rows is the key to increase the automation of field management in rice production. In order to avoid field management machinery rolling rice rows，machine vision and 2D LiDAR information were integrated to identify rice rows and perform navigation control of rice row tracking. Firstly, the rice row centroids were extracted from machine vision and LiDAR respectively, and the spatial coordinates and target areas were unified, and then a robust regression algorithm was used to fit the rice row centroids to obtain the navigation baseline and calculate the navigation parameters. Then a pre-sight tracking PID controller was designed. Finally, a rice row tracking and navigation test platform was built and experimental studies were conducted. The test results showed that the standard deviation of curve navigation test tracking simulated rice rows was 27.51mm;the standard variance of lateral deviation of rice rows navigation test tracking mechanical shift was 43.03mm and the standard variance of heading deviation was 3.38°.