Under the framework of total factor production, this study defined the concept of water resources input congestion. Based on the FGL (F?re, Grosskopf, and Lovell) model, this study identified the “congestion” of water resources input in 31 provinces of China’s mainland from 2000 to 2016, separated the inefficiency caused by water resources input congestion from the overall inefficiency and clarified the root causes of inefficiency. The results show that：(1) The congestion of water resources input is widespread in China. During the study period, the congestion degree of the provinces changed from low to high and agglomeration to dispersed, and the spatial differences of the congestion degree of water resources input among the provinces were significant. Among them, Beijing, Zhejiang, Guangdong and other provinces and cities had low congestion degree. On the whole, the central and western regions with unbalanced input structure was more prone to congestion than the eastern regions with high level of economic development. (2) The overall inefficiency of water resources utilization in Chinese provinces presented different trends. Most provinces were jointly driven by congestion inefficiency and pure technology inefficiency. From the inter-provincial and regional perspective, the ineffectiveness rate of pure technology was the dominant factor. Inefficiency of congestion was also an important factor affecting the overall inefficiency. The government should clarify the root causes of inefficiency, adjust the input factor structure in a timely manner based on local conditions, propose improvement directions for water resources utilization, and improve water use efficiency in order to promote sustainable economic development.