Based on traditional satellite attitude determination simulator, the paper studies a state estimation method to determine the satellite attitude, which is simulated in computer. This simulator is mainly composed by two parts: sensor model library and determination algorithm library. The sensor model library contains several sensor models which are implemented in C/C++, so as the determination algorithms library. In order to realize high accuracy measurement, hybrid extended Kalman filter(HEKF) has been used to estimate attitude parameters in the combination of gyro and star sensors. The simulation results show that HEKF can estimate high accuracy attitude parameters, the estimation errors of angles and angular rates are at 10-2 (° or °/s) level, and the estimation errors of gyro drifts can be in 10-3 (°/s) level.