The highly efficient linear and elliptical antenna array design presented in this article is intended to synthesize far-field radiation patterns for 5G communication. The objective is to achieve the best possible radiation pattern, which includes a lower side lobe level (SLL) and a narrower half-power beam width (HPBW). A low SLL is crucial to minimize intervention in the entire side lobe region, whereas long-distance communication requires a low HPBW. The marine predator’s algorithm (MPA) is used here to determine the optimal feeding currents to each array element to reduce the SLL and HPBW values. Two linear antenna array (LAA) design examples and the three elliptical antenna array (EAA) design examples are discussed in this article which are achieved using the optimal current excitations to each array element. Two well-established evolutionary optimization techniques, namely the genetic algorithm (GA) and the differential evolution (DE) technique, are also employed for comparison to demonstrate the superiority of the MPA optimization-based technique. The results achieved utilizing the MPA show the improvement of SLL contraction compared to the uniform antenna array and the methods affirmed in the recently published article.