The sports and recreation industry is a crucial part of a nation's economic development. In this paper, a reconfigurable multi-sensor water quality monitoring system based on Grey-ANFIS learning method is explored. Correlation of water quality factors is implemented as fuzzy inference. This proposed method provides an effective way to eliminate the trial-and-error process for finding the fuzzy rules and to predict the water quality evolvement appropriately. This project implementation is built on the ARM-based embedded Linux system which creates a low cost, high performance and easy development platform. The result of the proposed design method using intelligent agent is also given to illustrate the effectiveness in practice. It is expected to provide new ideas for academic industry cooperation.