Non-point source pollution in the planting industry is one of the important sources of water pollution, and the fragmentation of cultivated land will have a certain impact on it. Reasonably calculating the amount of non-point source pollution in planting and fragmentation of cultivated land and analyzing the correlation between the two is of great significance for protecting the agricultural water environment. Based on multi-period land use grid data from 2010 to 2021, this article used the Inventory Analysis and Equal Standard Pollution Load Method to analyze the sources of multi-scale non-point source pollution in planting in the Yangtze River Economic Belt from the perspective of region, province, and city. Then, ArcGIS and landscape pattern software Fragstats were used to calculate the multi-scale fragmentation of cultivated land in the Yangtze River Economic Belt. Finally, Pearson Correlation Analysis and Grey Correlation Analysis were used to deeply analyze the multi-scale correlation between non-point source pollution in planting and fragmentation of cultivated land. The research results indicated that: (1) At the regional scale, the main pollution unit of non-point source pollution in planting of the Yangtze River Economic Belt was cropland (441500 tons), and the main pollutant was total nitrogen. At the provincial level, the non-point source pollution in planting was mainly concentrated in Hubei Province, Anhui Province, and Jiangsu Province. At the city level, 28.13% of prefecture-level cities had an average non-point source pollution of over 5000 tons in planting. (2) At the regional scale, the Yangtze River Economic Belt has gone through a stage of “Increasing fragmentation-Fluctuating fragmentation”. At the provincial level, Guizhou Province, Shanghai City, Yunnan Province, and Chongqing City have all achieved higher fragmentation. At the city scale, 40.77% of prefecture-level cities were in a state of higher fragmentation and highest fragmentation. (3) At the regional scale, the spatial characteristics and clustering degree of cultivated land had a positive impact on non-point source pollution in planting. At the provincial level, the top three factors affecting non-point source pollution in planting in 11 provinces and cities were the Patch Cohesion Index (COHESION), Perimeter Area Fractal Dimension (PAFRAC), and Area-weighted Mean Shape Index (AWMSI). At the city scale, the correlation between the Interspersion and Juxtaposition Index (IJI) and non-point source pollution in planting in Hangzhou, Huzhou, and Jiaxing was relatively higher. The research results can provide reference for the scientific management of non-point source pollution in planting and the formulation of refined cultivated land protection policies.