Monitoring total suspended solids (TSS) in lakes at a global scale is critical for understanding lake ecosystem responses to climate change and anthropogenic activities. However, reliable retrieval methods for TSS global mapping remain elusive due to the optical complexity of inland waters, leaving the spatiotemporal dynamics of global lake TSS poorly constrained. This study developed a global TSS retrieval model using a Random Forest (RF) algorithm based on Landsat OLI surface reflectance imagery. The model achieved satisfactory performance (R-2 = 0.73) and demonstrated exceptional robustness in estimating lake TSS concentrations (0-1,500 mg L--(1)) across diverse water quality conditions, geographical locations, and temporal ranges. During summers of 2014 to 2023, lakes in mid-low latitude regions, particularly arid zones, exhibited significantly higher mean TSS concentrations compared to other areas. Our analysis revealed that 62.9% of clear lakes (TSS < 10 mg L--(1)) that underwent statistically significant changes (P < 0.05) showed significantly increasing trends of TSS, emphasizing the urgent need to enhance monitoring and protection measures for clear-water ecosystems. Furthermore, we classified lakes into three types based on the predominant regulatory mechanisms controlling TSS dynamics and systematically elucidated the distinct mechanisms through which lake topography, hydrological conditions, climate, and watershed vegetation cover influence TSS concentrations across different lake types. This study provides new insights into the spatiotemporal patterns of global lake TSS variations and the response mechanisms of different lake types to hydrological conditions and climatic forcing, contributing improved understanding of global freshwater ecosystem dynamics under environmental change.