Global Horizontal Irradiance (GHI) is a critical determinant for solar energy applications, since it directly influences the performance and efficacy of photovoltaic systems. Accurate forecasting of GHI is essential for optimizing energy production, integrating with the power grid, and formulating efficient energy management strategies. Traditional methods for forecasting GHI, such as physical models, sometimes struggle with substantial variations and complex meteorological conditions. Conversely, machine learning algorithms have emerged as viable instruments for GHI forecasting by using huge datasets and deriving insights from historical patterns. This research examines the use of several machine learning techniques, namely Gaussian Process Regression (GPR) and support vector machines (SVM), for forecasting the GHI in Dehradun city based on performance criteria. Our aim is to identify the most effective approach for capturing temporal dependencies and nonlinear correlations in GHI data via the assessment of various models’ performance. GPR outperforms SVM, NN, and DT, achieving R2, RMSE, and MAE values of 0.92286, 71.3320, and 47.6920, respectively. The research includes a comprehensive assessment of the model's accuracy, computational speed, and the impact of input variables, including meteorological parameters and historical GHI values.