In recent years, the risk assessment of changes in land use with respect to their ecological impact is still an unsolved problem. The water quality of rivers is influenced by the natural environment and human activities in the river basin. The latter of these two factors, especially change in land use, is the major factor in the deterioration of water quality. Therefore, the management of river pollution and water protection has significant academic and practical implications in terms of investigating any change in water quality relating to aspects of land use and changes in its appearance. We conducted a case study in the Liu Xi River watershed in Guangzhou city to examine the correlation between watershed landscape pattern characteristics and water variables. The basin was divided into 27 sub-watersheds and water samples were collected for monitoring of water quality. Water variables include COD_Cr, TP, NH₃-N and NO₃-N+NO₂-N. The values of these variables represent the degree of eutrophication and the content of organic matter, which are the characteristics of deteriorated water. We believe that a decrease in these variables results from the impact of human activities on water bodies. The choice of landscape indices was based on previous research, and included three scales. We used Fragstas3.3 software to obtain landscape indices of every watershed, then performed a correlation analysis. Spearman correlation and redundancy analysis (RDA) methods were used to analyze the relationship between water quality and land use patterns. Correlation analysis was made at buffer zone and sub-basins using three spatial scales. The buffer zone included a 30 m buffer and a 61 m buffer. Results showed that the land use structure of the Liu Xi River watershed had a great impact on water quality parameters. Residential land strongly affected water quality, and was the strongest impact factor for river pollution in the Liu Xi River watershed. In other words, if residential land increases, then the water quality will worsen. Forest constitutes a natural water purification system, so forestland exhibited a significant negative correlation with water quality. Orchards play a complex role in water quality, with an uncertain correlation between orchard and water variables. From upstream to downstream, urbanization gradually increased the watershed landscape features; the water quality also changed in response to this gradient. Water quality upstream is better than water quality downstream. The change in land use and land management practices caused by human activities and urbanization changes had a significant influence on water quality. The former brings more contaminants into the surrounding environment, while the latter increases the amount of impervious surface area. The results showed a strong correlation between water quality and landscape indexes. Landscape fragmentation and water quality showed a significant positive correlation, while landscape aggregation and patch shape complexity were negatively correlated; the differences between riparian and sub-watershed scale landscape spatial characteristics' effects on water quality were not obvious. The result of RDA revealed that water quality showed an obvious trend with the varying area percentage and edge density of the residential land. The correlation between water quality and landscape structure is insignificant with both scales. Based on our survey, land use change is a consequence of human activities, and changes the ecosystem status of the river and watershed. This could explain why water quality varies as the watershed landscape structure changes.