Aiming at the problem that the original output information of the micro inertial measurement unit is affected by the errors of bias, scale factor and nonorthogonal error, a fast insitu calibration scheme of MEMS IMU without highprecision turntable is proposed in this paper. Based on the analysis of the output characteristics of MEMS inertial sensors, the sensor error model is established. The calibration scheme of IMU 24-position is designed by using hexahedron fixture. The sensor error calibration is completed by using the gravity and each rotation angle as reference information. A calibration model is constructed for 9 parameters of accelerometer bias, scale factor and nonorthogonal error. The Newton method is used to estimate the optimal value of error parameters. Considering the 6 error parameters including gyro bias and scale factor, the optimal estimation of error parameters is calculated by least square method. The calibration compensation experiments on the accelerometer and gyro are carried out respectively. The experimental results show that the proposed fast insitu calibration method for the MEMS IMU can quickly obtain the sensor error parameters and improve the output accuracy of the data.