The paper presented an idea of applying information fusion technology based on RBF network in fault diagnosis of gas sensor. Through fusion of related information that influenced gas density of observed point, and using the comparison between the difference output of high-precision RBF network approximator and actual output of gas sensor and the given threshold,this technology realized monitoring and diagnosis for fault of gas sensor. Experiment showed that the state monitoring and fault diagnosis of gas sensor could be done effectually by this technology.