The unscented Kalman filter (UKF) became very attractive for the navigation sensors data fusion, because of algorithm significant accurancy and implementation advantages. The unsected Kalman filter is based on the unsected transform (UT) to perform the estimation of the system states. The main idea of the unsected transformation is following. It`s more effective to approximate probility distrubution function than arbitry transformation or nonlinear funcion. The developed sensors data fusion algorithm using the UKF is considered in this work. This algorith was applied for the state estimation of the loosely coupled GPS/INS integrated navigation system.