To determine the variations of nutrients, the observation data on water quality are analyzed at five monitoring sites of Baihua Reservoir from January 2014 to July 2018, including the characteristics of annual variation of dissolved oxygen, permanganate index, ammonia nitrogen, total phosphorus (TP) and total nitrogen (TN). The Bayesian theory based on entropy weight method was used to evaluate the comprehensive water quality of the five monitoring sites according to the situation of Baihua Reservoir. The driving force analysis on the water quality variation is carried out in terms of ecological restoration, industrial pollution sources, nutrients in sediments, seasonal thermal stratification of reservoirs and agricultural and domestic pollution sources. The results show that during the period of 2014 to 2018, the water environment of Baihua Reservoir was fluctuated greatly. Except TN, the concentration of each pollution index decreased with different degrees. The comprehensive water quality posterior probability indicates that the water quality of Baihua Reservoir had gradually deteriorated and the governance of TN and TP should be the main manner of pollution controls. The key driving factors of the water quality were getting better due to the development of ecological restoration projects and the reduction of industrial pollution sources, while the increase of agricultural and domestic pollutions were the main factors for the deterioration of water quality in Baihua Reservoir.