The nature of Photovoltaic (PV) power generation has a major impact on existing power systems, even though solar power has rapidly grown in importance as an energy source in many nations in recent years. For this an accurate solar power forecasting techniques are necessary to lower this uncertainty and preserve system security. This paper proposes, a novel forecasting strategy that combines Attention‑Gated Deep Long Short Term Memory (AG-DLSTM) and Salp Swarm Algorithm (SSA). The system involves, utilizing data pre-processing involves data cleaning method for findings the missing data and shows the duplicate values. Also, data transformation utilized for transformer raw data’s into statistical format. By data exploration univariate and bivariate analysis for PV power forecasting. The univariate analysis have single variable to analyse and bivariate analysis have two data’s to analyse. Also, feature scaling using mini-max scalar, and captured the maximum and minimum feature data. Finally, the proposed Deep Salp Swarm Algorithm based Attention Long Short Term Memory (DS2-ALSTM) for accurately predicting PV power .Implementing Python software, the proposed AG-DLSTM based SSA accommodate the actual generation pattern better than existing methods and produce the Mean Absolute Error (MAE), Mean Square Error (MSR) and Root Mean Square Error (RMSE) value of 0.53, 0.28 and 0.43 respectively.