This paper introduces and evaluates novel hybrid deep learning models for solar power forecasting using time series data. The study investigates the effectiveness of various models, including combinations of convolutional neural networks (CNN), long short-term memory (LSTM), and transformer (TF) models. The hybrid CNN–LSTM–TF model outperforms other models, achieving a mean absolute error (MAE) of 0.551% when using the Nadam optimizer. In contrast, the TF–LSTM model performs poorly with an MAE of 16.17%, highlighting the challenges in accurately predicting solar power. The research demonstrates the importance of selecting appropriate models and optimizers for accurate solar power forecasting. This is the first study to comprehensively explore transformer networks in hybrid models for solar power forecasting. The findings contribute to optimizing renewable energy systems and improving grid stability and energy management. The study also highlights the benefits of hybrid models in capturing complex patterns and dependencies in solar power data, leading to more accurate predictions. The results show that hybrid models, such as CNN–LSTM–TF, significantly improve forecasting accuracy compared to single models. The study emphasizes the importance of model selection and optimization in achieving reliable solar power forecasts, which are crucial for effective renewable energy integration and sustainable energy planning.This paper introduces and evaluates novel hybrid deep learning models for solar power forecasting using time series data. The study investigates the effectiveness of various models, including combinations of convolutional neural networks (CNN), long short-term memory (LSTM), and transformer (TF) models. The hybrid CNN–LSTM–TF model outperforms other models, achieving a mean absolute error (MAE) of 0.551% when using the Nadam optimizer. In contrast, the TF–LSTM model performs poorly with an MAE of 16.17%, highlighting the challenges in accurately predicting solar power. The research demonstrates the importance of selecting appropriate models and optimizers for accurate solar power forecasting. This is the first study to comprehensively explore transformer networks in hybrid models for solar power forecasting. The findings contribute to optimizing renewable energy systems and improving grid stability and energy management. The study also highlights the benefits of hybrid models in capturing complex patterns and dependencies in solar power data, leading to more accurate predictions. The results show that hybrid models, such as CNN–LSTM–TF, significantly improve forecasting accuracy compared to single models. The study emphasizes the importance of model selection and optimization in achieving reliable solar power forecasts, which are crucial for effective renewable energy integration and sustainable energy planning.