Abstract: Redundant design is an important method to increase reliability of small satellite. Traditional redundant design methods in small satellite design use spare hardware, therefore greatly increases the power consumption, volume and weight of the satellite. A redundant design method which combines software and hardware is presented in this paper. In this method sensors are divided into several groups. Multi sensor data fusion algorithm based on Bayes rule is used to give the estimate value for each sensor in a group. The validity of the estimate value is judged. If the estimate is invalid, the spare sensor is enabled. This redundant design method uses software to judge the validity of the result of temperature measurement and hardware to provide spare sensor. Therefore the number of spare sensors can be reduced significantly.