The pavement performance of superficial soils in southern Nigeria is in decline, most notably due to the use of inferior pavement materials and poor soil conditions. The study, therefore, examines the impact of sand, cement, and binary (sand/cement) stabilisation on soils in an attempt to improve road development in the City of Warri. The natural condition of sixteen (16) soil samples was subjected to consistency tests, classification tests, compaction, and soaked California Bearing Ratio (CBR) analy-ses as a standard for comparative analysis. The effects of sand, cement, and binary stabilisation on the engineering behaviour and suitability of natural soils were examined using specified road stan-dards, CBR responses, and t-test analysis as bases for comparison with the natural, untreated soils. The properties of natural and treated soils were fed into machine learning predictive model, including Random Forest Model (RFM), Extreme Gradient Boosting (XGBoost) Model, and Explainable Boosting Model (EBM), for CBR prediction evaluation. Results confirmed that sand stabilisation had a significant impact on the soil grade, with a marginal impact on the natural CBR (resulting in a slight boost from the natural range of 3-17.9% to 11-28.8%), while the soil suitability remained constant. Cement-treated soils were improved to subbase/base quality (71.3-193.3%) at 7 wt. % cement. Binary stabilisation resulted in base-quality soils (107-272.5%). T-test analyses confirmed that binary stabilisation is the most tech-nically viable solution to the pavement deficiency in the City. The study presents a feasibility resource database for the stabilisation of deficient superficial soils for road development in deltaic environments.
