To capture equipment fault in real time and automate fault diagnosis, a pattern recognition method, based on data eigenvector and TCP transport protocol, was proposed to capture Water Quality Monitoring equipment’s fault information. Fault data eigenvector was designed after analyzing the equipment fault feature and capture strategy, structural pattern recognition strategy was confirmed and specific data frame was designed in response to the fault data eigenvector, by integrating the data frame design into the equipment’s communication protocol, data related to different fault compiled into fault data frames by transmitters or communication module of equipment’s different components, the remote sever captures equipment fault on transport via fault data frames according to the structural pattern recognition strategy. With 7 months of practical application in Taihu aquaculture project and research center of agricultural information technology, combining with historical fault data and contrast with artificial recognition result, the simulate experiment shows this method has higher response rate and process rate with a nice accurate.