Enhancement of performance of antenna array in terms of directive characteristics requires simultaneous control of both side-lobe level (SLL) and beam width (BW). The population-based evolutionary computing techniques are best suited for such array synthesis problems. In this paper, a novel evolutionary computing tool known as flower pollination algorithm (FPA) is applied to linear array synthesis problem. In addition, the performance of both amplitude only and amplitude–spacing-based methods is assessed. The proposed method estimates the weights of each objective so that the generated radiation patterns with desired SLL and BW are produced. The performance and the efficiency of FPA-based method are evaluated and also compared with those obtained using genetic algorithm and particle swarm optimization. The problem of synthesizing patterns with narrow and wide nulls is also investigated. Multiple narrow band nulls as well wide nulls are produced using the FPA with amplitude only method. The investigation has been made for both sum and difference patterns.