The effect of utilising energy derived from fossil sources on the environment has aroused research interest in alternative and sustainable energy sources. Synthesis gas, a mixture of carbon monoxide (CO) and hydrogen can be used as starting materials in hydrogenation reactions to produce chemical intermediates that can be used in various processes. This study investigates the robustness of applying a fine Gaussian support vector machine algorithm for predicting light olefins from catalytic CO hydrogenation using magnesia nanoparticles-based catalysts. The datasets obtained from the CO hydrogenation reaction consist of input parameters such as magnesia nanoparticles contents, reaction temperature, and reactor pressure, and the output parameters which include CO conversions and the selectivity of light olefins (CH4, C2H6, C2H4 C3H8, C4H8, and C3H6). The dataset was trained and employed for the prediction of the light olefins using a support vector machine with an inbuilt Fine Gaussian Kernel function. The performance of the support vector machine was evaluated using the coefficient of determination (R2), root mean squared error (RMSE), mean square error (MSE), and the mean absolute error (MAE). The support vector machine showed significant potential in the prediction of CO conversion, CH4 selectivity, C2H6 selectivity, and C2H4 selectivity as indicated by R2 of 0.770, 0.800, 0.730, and 0.930, respectively. While less predictive performance was obtained for the prediction of C3H8 selectivity, C4H8 selectivity and C3H6 selectivity as indicated by R2 of 0.630, 0.610, and 0.320, respectively.