This paper conducts research on storage testing and strapdown inertial navigation techniques base on MEMS IMU. A multi-channel, high-sampling-rate, micro-miniature and high-capacity storage testing device is designed to cope with the high-frequency and high-precision testing for MEMS IMU. An optimized data computing method is proposed for the strapdown inertial navigation base on the collected sensor data. When transplanting the algorithm, DSP functions are called to make full use of hardware resources and obtain higher data update frequency. A sine wave signal is generated to analyze the gain accuracy of the device and ideal IMU data with Gaussian white noise is fed to the device to quantitatively analyze the computational error of it. Experimental results show that the sampling frequency is around 1MHz, the relative error of gain is 1% or less, and the errors of algorithm are far less than the instrumental errors. While guaranteeing this performance, the device can perform long sampling periods of more than one hour and store 32GB of data. The performance of the system meets the expected design specifications for projectile-borne testing.