The sensitive point of inertial measurement unit does not coincide with the origin of the vector coordinate system, there exists angular motion in the process of calibration and application, and g the leverarm effect error is introduced. In order to solve this problem, this paper puts forward the leverarm calibration method based on Kalman filtering algorithm. According to the leverarm effect error model, the paper analyzes the connection between inside and outside the leverarm effects. By using the linear velocity error as the observed quantity, the Kalman filtering equation is established. Through analyzing the system observability, the error of the two axes linkage type incentives mode is designed. Using five swaying sea condition experiments as an example, the linear velocity error is reduced by 73% after compensating leverarm effect.