<p>Lack of accurate sediment yield data and efficient sediment yield prediction methodology have contributed to poor planning and designing of reservoirs, which has led to increased reservoir sedimentation and loss of storage capacity. The estimation of suspended sediment yield in ungauged catchments is imperative, yet extant models have the potential to be more efficacious in predicting the suspended sediment yield. Furthermore, the scarcity of data represents a pervasive challenge in the field of sediment yield prediction. The present study, therefore, sought to develop an alternative empirical model using the Quantum Geography Information System Soil Water Assessment Tool Plus (QSWAT⁺) conceptual model for a situation where data is scarce. The QSWAT⁺ model was used to divide the Kessie watershed into 18 sub-basins to determine the characteristics of each catchment and the basis for analyzing input variables. Data from six watersheds were collected over 11 years to test the performance of the newly established alternative model, which was assessed using model assessment statistics. Principal component analysis (PCA) was used to determine the significant variables that affected sediment yield. Through this analysis, seven significant variables were found among the 24 variables considered; the most significant included drainage area, slope of the stream, length of the main channel, rainfall, agricultural area coverage, forest area coverage, and streamflow. In general, climatic, geomorphological, and hydrological variables were found to be the key determinants of suspended sediment yield. The calibration and validation results demonstrated that six watersheds such as Robigumero, Gumara, Gilgel Abbay, Ribb, Kessie, Koga, and Megech achieved the following Nash-Sutcliffe Efficiency (NSE) values: (0.7, 0.7), (0.73, 0.69), (0.8, 0.73), (0.85, 0.7), (0.79, 0.7), (0.85, 0.8), and (0.8, 0.82), respectively. Following the calibration and validation of streamflow and suspended sediment yield data at the Kessie gauge station, a new empirical model was developed to estimate suspended sediment yield in ungauged catchments. The simulated suspended sediment yield estimated by the alternative empirical model aligned well with the observed values across all six watersheds. The model exhibited strong performance, with specific examples including Gumara (R² = 0.84), Ribb (R² = 0.71), and G. Abbay (R² = 0.61). The model’s validity was further confirmed using 11 years of data, demonstrating its reliability and accuracy. This improves the model’s accuracy and applicability in a variety of environmental scenarios, as well as its calibration and validation in places with little data. This approach could lead to better sediment management strategies and better planning for watershed conservation efforts. It is recommended that the model be evaluated in more basins, particularly with various soil properties.</p>