This study aims to develop models for estimating topsoil properties by analyzing both the parametric and textural features extracted from Sentinel-1 C-SAR (VV, VH) images. Field measurements were collected from 13 soil samples in the Laylan region of Kirkuk City, Iraq, and utilized to develop and validate the models. The study employed classification algorithms, including the Random Forest (RF) and Maximum Likelihood (ML) classifiers, using specific indicators derived from Sentinel-1 data. Additionally, a soil triangle was constructed using three axes to represent the predicted target parameters, facilitating the identification of five distinct soil groups in the study area. The findings reveal that the soil triangle enables the delineation of five subcategories of soil characterized by varying proportions of sand and silt. Each soil sample was categorized into one of five predefined classes based on its clay content, ranging from 0% to 14.48%. The performances of the ML and RF algorithms were assessed, demonstrating their effectiveness in estimating percentage labels despite limited training data, with ML exhibiting higher accuracy than RF. The developed models showed promising potential; however, their applicability should be tested across diverse geographic regions with varying climatic conditions. Future research could focus on utilizing these models to generate soil texture maps, potentially enhancing soil parameter estimation in different environments.