This paper proposed an Engineering Implementation scheme on fault diagnosis for a real quadrotor Unmanned Aerial Vehicle (UAV) with actuators fault. A novel nonlinear observer is utilized to estimate the state of the system and to generate fault offset residuals for fault diagnosis. The Fault Diagnosis (FD) rules based on residuals are designed to isolate and estimate faults. Unlike former research work in FD for quadtrotors, the proposed FD is based on real flying data but not model simulation data and still have an acceptable FD accuracy. Experiments for the rotor fault have been carried out to show the performance and effectiveness of the proposed observer method and the results are discussed.