This paper puts forward a real-time smart fault diagnosis system (SFDS) intended for high-speed protection of power system transmission lines. This system is based on advanced signal processing techniques, traveling wave theory results, and machine learning algorithms. The simulation results show that the SFDS can provide an accurate internal/external fault discrimination, fault inception time estimation, fault type identification, and fault location. This paper presents also the hardware requirements and software implementation of the SFDS.