Utilizing renewable energy sources is critical for achieving a sustainable future because it allows us to tap into natural resources in a way that maintains a long-term energy supply while avoiding depletion of the Earth’s resources. Worldwide, people extensively utilize the solar photovoltaic panel system as a sustainable energy source, offering a viable alternative to fossil fuel-based energy generation. Nevertheless, its primary limitation resides in its lack of predictability. Hence, it is crucial to establish a prognostication system to anticipate the energy production of solar photovoltaic systems. This prediction is based on multiple parameters, including solar radiation, humidity, and other variables. This study presents a machine learning prediction method for solar photovoltaic systems that incorporates three robust machine learning techniques: Extra randomized tree regressor, extreme gradient boosting, and K-nearest neighbors. In addition, the authors in this paper utilize hyper-parameter tuning to determine the most suitable settings and conduct a comprehensive analysis of the impact of seasonal data on prediction performance. The results of our study indicate that the Extra Tree regressor demonstrates superior accuracy compared to other models. Additionally, they found that seasonal data, especially humidity, significantly influences the performance of all machine learning algorithms.