Abstract: Since the birth, the FOG has been put into application in many ways, for its many advantages such as simple structure, stability, low cost, small volume and low weight. But with the effect of much noise, the stability of the FOG is still not good. In order to improve the stability, the data of IFOG are analyzed with Allan variance method and each error coefficient is identified. Furthermore, a random drift error model for IFOG was built by the aid of the time sequence analysis and the error model of the FOG′s drift is expressed as ARMA(2,1). At last, onestep predict curve of data is assessed. The adaptability of the error model is discussed. The conclusion provides supports for error compensation of IFOG.