The research status of systematic calibration was analyzed, and an inertial navigation system error model was established. Additionally considering the influences of quadratic term error coefficient of the accelerometer triad and the inner lever arm parameters to the system, a systematic calibration method was proposed based on 36-dimension Kalman filter. An appropriate calibration path was designed and a Kalman filtering model was established. Simulation and experiment results indicate that the accuracy of the estimated biases is better than 0.001(°)/h and 9 μg for the laser gyroscopes and the accelerometers respectively, the accuracy of estimated scale factor errors is better than 3 ppm and 2 ppm respectively, the accuracy of estimated installation error angle is better than 1″ and 3″ respectively, the accuracy of estimated quadratic term error coefficient is better than 4×10^-10 s²/m, and the accuracy of estimated inner lever arm parameters is better than 3 mm. Therefore, this method satisfies the demands of high precision inertial navigation system calibration.