Fuzzy basis function network (FBFN) based on T-S fuzzy model is given. A general approach for fault information detection in unknown systems using FBFN is present. The unknown system is composed of known part and unknown part. The output of an actual system is composed of three portions: the output of a mathematical model, the output of unknown part and fault information. The known part can be represented by a mathematical model. The unknown part, which includes the uncertainty of model error, disturbance inputs, etc, is estimated by a FBFN. The fault information in the unknown system can be estimated using the outputs of actual system, the mathematical model and FBFN.A simulation example of fault information detection in a microwave landing system of an aircraft is given.