Abstract: A fine alignment equation of SINS for a stationary base is derived.Then，an equivalence operation between fine alignment model and mechanization is expatiated and observability analyzed.We propose a key technology for a simplified fine alignment model，which improves the self-alignment Kalman filter model through three strategies;eliminating coriolis force and the unobservable state，transforming the bias in h frame to equivalent bias in n frame.Test results show that a simplified model is optimized as a fine alignment method for a stationary base.It makes the model simpler as well as computation more efficient;meanwhile，fast convergence and high estimation accuracy remain unchanged.