Abstract: In order to improve adaptability of autopilot system to influential factors such as longitudinal speed of agricultural vehicle and slip condition，a heading estimation algorithm based on slip rate estimation was proposed.Firstly，NF-752 crawler tractor was adaptively modified to build a tractor autopilot system test platform，its manual steering mechanism and stepless speed change mechanism were transformed into electrical signal controled by high torque steering engines，vehicle location and body speed were measured by RTK-GNSS positioning system.Secondly，based on vehicle speed and crawler speed on both sides as state variables，course control model was established considering crawler slip rate，and then a path tracking control method was proposed based on this model.Finally，straight path tracking tests were carried out on asphalt pavement at low speed（1 km/h），medium speed（5 km/h）and high speed（9 km/h），results showed that there was no significant difference in path tracking error under different operating speed conditions，and maximum path tracking error was -2.00 cm，standard deviation was 0.93cm.Curve path tracking experiments were carried out in experimental field，and results showed that when tractor tracked curve path at 6 km / h，tracking error was better than 10 cm，and there was no obvious error increase in slip area.Experimental results showed that proposed method had good adaptability to operation speed，and overcome influence of slip phenomenon on control precision.