The stability of the power sector has become uncertain due to the unpredictable characteristics of renewable energy sources such as solar photovoltaic (PV) power generation. It endangers the balance of the power system which is very sensitive to any mode of change and results in an ineffectiveness to match power consumption and production. The ultimate goal of harvesting renewable energy is to integrate it into the power grid. So, predicting the total amount of power generation by solar cells has become an important aspect. This study delineates various Convolutional Neural Network (CNN) techniques such as regular CNN, multi-headed CNN, and CNN-LSTM (CNN Long Short-Term Memory) which employs sliding window algorithm and other feature extraction and pre-processing techniques to make accurate predictions. Meteorological parameters such as Solar Irradiance, Air Temperature, Humidity, Wind Direction, and Wind Speed are related to the output of the solar panels. For instance, input parameters were taken for 5 years span and predicted for a particular day and one week. The results were evaluated by comparing them with traditional forecasting techniques such as Autoregressive Moving Average (ARMA) and Multiple Linear Regression (MLR). The efficacy of the result was also evaluated by the Evaluation Metrics such as RMSE, MAE, and MBE. Both traditional and machine learning techniques demonstrate the effectiveness in producing short-term and medium-term forecasting.