In this paper, an improved variant of nature-inspired meta-heuristic algorithm inspired by the pollination process of flowering plants called improved flower pollination algorithm (IFPA) is utilized for solving the optimal allocation of capacitor banks (CBs) and distribution-static synchronous compensator (DSTATCOM) problem considering electric vehicle (EV) load growth. In IFPA, a new double-direction learning strategy to advance local searching capacity, a novel dynamic switching probability method to balance global and local searching, and a new greedy technique to increase population diversity. A multi-objective function is formulated for minimizing the real power loss and installation cost of CBs/DSTATCOM. The search space of the multiple CBs/DSTATCOMs is primarily reduced using voltage stability index (VSI) and later the best locations and sizes of CBs/DSTATCOMs are determined by implementing IFPA. The proposed hybrid VSI-FPA approach is applied to solve the DSTATCOM allocation problem in standard IEEE 33-bus radial distribution systems (RDS). The effectiveness of the proposed approach is compared with the similar types of heuristic approaches in the literature. The comparative results shown that the IFPA is outperformed than other algorithms by providing minimum losses, reduced installation cost, and consequently improved voltage profile as well as voltage stability irrespective of EV load growth via allocating the CBs/DSTATCOM optimally in the RDS. The basic network profited from CBs and DSTATCOMs by a loss reduction of 34.71% and 29.50%, respectively, according to the results. The loss, on the other hand, increases to 88.69 percent when the network is filled with 50% EV load. With CBs and DSTATCOMs, however, the higher losses are just 37.07% and 37.26%, respectively.