Owing to its potential in offering a sustainable future, the renewable energy conversion system has gained considerable interest recently. Solar collector (water/air) has been increasingly used in the optimization of renewable solar energy and the equipment used in the process. The traditional solar air heater (SAH) system comprises a parallel plate and a solar radiation absorber plate, whereby the inlet air is heated, and acquired hot air is applied for a variety of applications. Despite the wide-ranging application, the traditional SAH has lower heat storage capacity and thus exhibits comparatively reduced thermal efficacy attributed to numerous factors. Subsequently, increasing efforts were being made on the overall air heater efficiency and the development of thermal performance. In this view, this work was focused on the design of Slime Mould Optimization with Multi-head Attention oriented two-directional Long Short-Term Memory Enabled Prediction (SMO-MABLSTM) model for thermal efficacy of Porous Bed SAH. The proposed SMO-MABLSTM model is utilized to predict the thermal efficiency of unidirectional flow porous bed SAH. In order to optimize the predictive outcomes of the MABLSTM model, the SMO algorithm is applied. The experimental outcomes verified that the SMO-MABLSTM technique is superior to other models in the predictive of SAH thermal efficiency.