Multivariate statistics have been applied to evaluate the water quality data collected at six monitoring stations in the Feitsui Reservoir watershed of Taipei, Taiwan. The objective is to evaluate the mutual correlations among the various water quality parameters to reveal the primary factors that affect reservoir water quality, and the differences among the various water quality parameters in the watershed. In this study, using water quality samples collected over a period of two and a half years will effectively raise the efficacy and reliability of the factor analysis results. This will be a valuable reference for managing water pollution in the watershed. Additionally, results obtained using the proposed theory and method to analyze and interpret statistical data must be examined to verify their similarity to field data collected on the stream geographical and geological characteristics, the physical and chemical phenomena of stream self-purification, and the stream hydrological phenomena. In this research, the water quality data has been collected over two and a half years so that sufficient sets of water quality data are available to increase the stability, effectiveness, and reliability of the final factor analysis results. These data sets can be valuable references for managing, regulating, and remediating water pollution in a reservoir watershed.