Modeling of physical properties and determination of the chemical composition of BTS were studied at 12.25% wet basis. Mass correlation with principal dimensions was modeled and evaluated using statistical parameters including chi-square ($\chi^2$), coefficient of determination (R$^2$), root mean square error (RMSE), and standard error (SE). Chemical properties were analyzed using standard techniques. Results showed mean values of seed mass, major, intermediate, and minor diameters of 0.18$\pm$0.02 g, 0.85$\pm$0.007 mm, 0.73$\pm$0.074 mm, and 0.38$\pm$0.046 mm, respectively. The seed was spherical (72.64%) with mean aspect ratio (1.17$\pm$0.124), ellipsoid ratio (2.27$\pm$0.31), and eccentricity (0.4768$\pm$0.16). The correlation of mass with intermediate diameter, and with the interaction of intermediate and minor diameters, produced the best non-linear regression models. The intermediate diameter model yielded the highest R² (98%) and lowest $\chi^2$ (2.4 $\times$ 10$^{-5}$). The interaction model achieved R² of 99% with the lowest RMSE (0.0032), $\chi^2$ (0.0029), and SE (1.04 $\times$ 10$^{-5}$) for predicting seed mass. Proximate and phytochemical compositions were not significantly different ($p>0.05$) from recommended levels by WHO and NAFDAC. This study provides a basis for developing standardized, cost-effective processing methods to enhance market value and ensure compliance with international quality standards.