For developing countries, an adequate domestic water supply is conventionally assessed based on the proportion of communities that are covered by improved water sources. However, it is difficult to evaluate water poverty accurately, as it is multidimensional. For this reason, this paper used the Water Poverty Index (WPI) to measure water poverty in rural communities of arid areas in China. This study also uses the Least Square Error (LSE) model to analyze the influencing factors of water poverty. Based on the WPI and LSE, the results showed that the rural communities of Sheshu, Fanyao, Dongcao, Qiaodi, and Gouershang (listed in order of priority of need for intervention) are in a water poverty situation. In rural communities with high water poverty, the suggested priority order for the study was environment, capacity, use, resources, and access, with the environmental factor needing to be improved. The results are useful for prioritizing areas and identifying the extent of the need for policy intervention on different scales. The research findings are intended to complement the evaluation of water poverty and to provide a strategy for regional water resources management to relieve water poverty.