During guided retrofit of bombs, the adoption of MEMS-IMU (inertial measurement unit) can significantly cut the cost. However, insufficient accuracy is a big problem with MEMS-IMU. To solve the problem, kinetic pedestal transfer alignment is adopted to eliminate initial error and improve system accuracy. At first, the authors set up an error model of the main and subordinate inertial navigation system. Then, Kalman filter is used to estimate the measurement error of the MEMS-INS. Therefore, the error can be corrected in the gesture matrix, and the model can feed the MEMS-INS with the best initial value at the time of launch, which, in turn, improves system accuracy and makes it more feasible. In a word, kinetic pedestal transfer alignment can improve the accuracy of MEMS-IMU and the combination provides a low-cost, high-accuracy option for the guided retrofit of bombs.