Streamflow is one of the key components of the hydrological cycle in a watershed that can be affected by a variety of variables. One of the key variables is the change in land use and land cover (LULC). This study is mainly focused on assessing the impact of LULC change on only surface runoff using the soil and water assessment tool (SWAT) model. Sensitivity, calibration, and validation were performed by SWAT-CUP (calibration and uncertainty program) using sequential uncertainty fitting version 2 (SUFI-2). The time series LULC map of the research area for the years 1997, 2010, and 2022 were assessed using Google Earth Engine (GEE). There were four different algorithms applied to classify the LULC change (SVM, CART, RF, and Navia’s) available in GEE and the coverage of LULCs was including land class of agricultural land, forest land, grass land, urban area, barre land, and shrub land. The best performing of four distinct algorithms were compared, and chosen to generate LULC map. To train and evaluate the LULC map generated from the GEE platform, high-resolution, 30 meter Landsat imagery from Landsat 5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper (ETM), and Sentinel 2 were employed, along with historical trends and ground-based data. As a result, the RF algorithms performed well. During the study period, the area covered by agriculture increased from 48.09% to 64.51%, while forest cover increased from 1.52% to 5.51% and grass land declined from 32.31% to 8.81%. The calibration and validation result showed that an acceptable range between observed and simulated streamflow (0.83 for NSE and 0.83 for R2) and (0.77 for NSE and 0.80 for R2) respectively. LULC changes have an impact on the streamflow of the Kiltie watershed by changing surface water quantity increased by 20,971,440 m3at wet and decreased by 3,090,528 m3 at dry season. Key words: SWAT, SWAT-CUP, Streamflow, Google Earth Engine, Machine Learning, Kiltie Watershed