Monitoring the Maximum Power Point Tracking (MPPT) is one among the most crucial jobs in a solar system. Significant issues with solar power generation, including the installation and operation of PV panels, include harmonic distortion and electromagnetic radiation interference. As a result, the MPPT controller in the proposed work utilizes deep learning to track the MPP. Developing a Deep Belief Network (DBN) using the generated dataset allows it to determine the MPPT. Utilizing the Green Anaconda Optimization (GAO) based DBN to provide the reference or attained voltage, the fractional order controller is then used to generate the converter’s pulse. The suggested methods outperform prior MPPT methods in terms of steady state response, oscillation-free operation, loss reduction, and time to settle down at MPP. The suggested model provides very low harmonic values, such as 0.78% for voltage THD and 0.781% for current THD, and accuracy of 99.8%. The outcomes show that, the optimal DBN MPPT approach works better than the controllers.