A dynamic neural networ k fuel measurement model is constructed firstly, related problems on how to use this model in reality is discussed followed, a simulated result is provided at t he end.