Water quality information collection is an important part of factory aquaculture. This paper proposes a kind of wireless sink nodes using information fusion for water quality information collection in factory aquaculture. In the sink nodes, Support Vector Regression and fuzzy algorithmic approach are used for information fusion. Making decisions according to information fusion, it converts the collected water quality information into a simple parameter that signified the current state of water quality. The sink nodes can eliminate redundant information, reduce information transmission, thus save energy effectively and prolong the network life.