In this paper, we propose a New Artificial Bee Colony NABC algorithm to enhance the search ability of onlooker bees. In NABC, the employed bees and onlooker bees utilise different search strategies to generate new candidate solutions. Moreover, NABC does not use the roulette selection, and a new method is designed to select good solutions for the onlooker bees. Experiments on some famous test functions show that the NABC algorithm outperforms some similar ABC and swarm intelligence algorithms.