Abstract: Water is the most important and sensitive factor in watershed spatial management, and the water sensitivity difference in different sections of the basin will have different restrictive effects on the expansion of construction land. From the perspective of integrated watershed spatial management, this paper attempts to construct watershed water sensitivity regionalization evaluation indexes and methods from four aspects: water resources, water ecology, water environment and water disasters. Taking Taihu Lake basin as a case, water sensitivity evaluation based on small watershed units is carried out. Based on the land use data of 1985, 1995, 2007 and 2017, the correlation between construction land expansion and water sensitivity in Taihu Lake basin in the past 30 years is depicted. The research shows that: ①From 1985 to 1995, the expansion of construction land in Taihu Lake Basin was basically not related to water sensitivity, and the construction land in high water sensitive area has the highest expansion intensity; ②From 1995 to 2007, the correlation between spatial expansion and water sensitivity was enhanced, the expansion intensity of construction land decreased with the increase of water sensitivity, and the area with larger expansion intensity was dominated by low water sensitivity area; ③From 2007 to 2017, the expansion intensity of construction land has been significantly improved. The expansion space is mainly distributed in low-water sensitive areas, and the restriction of water sensitivity on spatial expansion was increasing; ④Although the restriction of water sensitivity on the expansion of construction land is increasing,the correlation between them is not high enough, and the water sensitivity has not become a key constraint on construction land expansion. The research results of this paper can provide a scientific basis for coordinating the relationship between socio-economic development and resource environment in the basin and formulating spatial management policies for adapting to water sensitivity.