Suitability of the inertial navigation algorithm to the error model is the premise of the design of the static self-alignment or self-calibration Kalman filter. The open-loop navigation algorithm and the simplified Strap-Down Inertial Navigation System (SINS) static error model are given. The open-loop navigation algorithm is proved suitable to the simplified error model by error comparison. Based on the algorithm and model, a closed-loop Kalman filter is designed. The filter is used in semi-physical simulation of multi-position alignment, and the result shows that the estimation of gyroscope biases and accelerometer biases has the accuracy higher than 97%. The application proves further that the open-loop navigation algorithm is suitable to the simplified error model.