Simulating multi-scale dynamics of physical phenomena is complex and presents an ongoing challenge to computational science. While partial differential equations (PDEs) form the basis for most computational models of natural phenomena, their numerical approaches suffer due to high computational costs, work at a reduced speed, and make tradeoffs in accuracy. New developments suggest the effectiveness of neural networks of deep learning in solving PDEs. This study offers the construction of the Aquila Optimized Physics-Informed Neural Network (AO-PINN) in an attempt to effectively address the limitations discussed above. The AO-PINN incorporates Aquila Optimization (AO) to optimize the process of learning, minimize error, and increase the rate of convergence. This model employs sine functions to numerically solve PDEs and employs the physical information as regularization. Experimental results at a correlation of 0.8858, R2 (0.784), MAE (1.843), RMSE (2.56), and MSE (6.565) attest to its effectiveness. These outcomes demonstrate that AO-PINN minimizes error and training time, hence ensuring an efficient and accurate way of solving intricate PDEs.