The failure detection of the oil and gas pipeline corrosion involve many characteristics, such as randomness, complex, multiple factors, and non-linear so on, it is difficult to describe by using the precise mathematical model. This paper proposes a failure detection method on pipeline corrosion, which is the Fuzzy neural network, based on Hybrid the improved particle swarm algorithm. It optimizes the convergence of Particle swarm algorithm and accelerates the speed of finding the optimal solution, which is used in the building of the Fuzzy neural network, and in this way, it sets up a pipeline corrosion failure detection model. It obtains a good recognition effect, by using the actual pipeline corrosion failure detection data to diagnostic applications, and the result verify the feasibility and validity of the model and algorithm.