The paper presents the effectiveness of an intelligent control based Dynamic Voltage Restorer (DVR) for voltage related issues. The developed neural architecture optimized by Coronavirus Herd Immunity Algorithm (CHIA) to extract the best fitted feedforward (FF)-based Backpropagation (BP) which is employed for the fundamental extraction. The distribution system voltage is stabilized by the implementation of Fractional Order PID (FOPID) and effective regulates the voltage across DC and AC bus under dynamical state. The main disadvantage of classical technique SRF-PI has manual intervention for coefficient tuning and increases the computation effort. The proposed controllers FF-BP-based CHIA and optimized FOPID by Sooty Tern Optimization overcome the issues of SRF-PI and significantly improve the overall DVR performance. In the proposed design of DVR, NN layer interconnected weights and neurodes are optimized by Coronavirus Herd Immunity Algorithm (CHIA) and gains of FOPID by Sooty Tern Optimization (STO). The objective function taken for assessment of FOPID is the integral time square error (ITSE). The FOPID \((PI^{\lambda } D^{\delta } )\) has additional tuning coefficients compared to classical PI tuning and result confirms the less settle time, rise time, and overshoot of 0.223 s, 0.126 s, and 6.6% respectively. The controller efficacy is examined under polluted grid voltage like sag/swell, distortion, and imbalance. Compared to the conventional control scheme, FF-BP-based CHIA and optimized FOPID improves the output voltage profiles and maintain THD of 3.33% as per IEEE-519. The DVR performance is validated by simulation and experimental results with the developed control algorithms.