Abstract: In order to solve the problem that the small defect signal in the guided wave detection of the actual pipeline is inundated by the characteristic echo of the complex structure, an effective pre-processing method for the signal is proposed for suppressing the effect of those possible external factors. Theoretical study on the algorithm for residual signal analysis in the time domain is carried out, and the feasibility of extracting the characteristics of tiny defects is pointed out. A remote monitoring cloud platform for pipeline damage with magnetostrictive ultrasonic guided waves is built to realize automated, digital and intelligent pipeline online monitoring. The corrosion rate and health status of the pipeline are obtained and the feasibility and effectiveness of the monitoring algorithm and platform are verified by laboratorial experiments.