This paper proposes a new hybrid least squares support vector machine and artificial bee colony algorithm (ABC-LS-SVM) for multi-hour ahead forecasting of global solar radiation data. The framework performs on training the LS-SVM model by means of ABC using measured data. ABC is developed for free parameters optimization for LS-SVM model in a search space so as to boost the forecasting performance. The developed ABC-LS-SVM approach is verified on an hourly scale on a database of five years of measurements. The measured data was collected from 2013-2017 at the Applied Research Unit for Renewable Energy (URAER) in Ghardaia, south of Algeria. Several combinations of input data have been tested to model the desired output. Forecasting results of 12th hour ahead global solar radiation with ABC-LS-SVM model led to an RMSE error equal to 116.22 (Wh/m2) and correlation coefficient R2 = 94.3 (%), and RMSE=117.73 (Wh/m2) and correlation coefficient R2 = 92.42 (%) with classical LS-SVM. The achieved results reveal that the proposed hybridization scheme achieves high performance over the stand-alone LS-SVM.