In this paper, a novel approach, called Termite colony optimization (or TCO), for optimizing numerical functions is presented. TCO is a population based optimization technique which is inspired from intelligent behaviors of termites. The proposed approach provides a decision making model which is used by termites to adjust their movement trajectories. Termites move randomly in the search space, but their trajectories are biased towards regions with more pheromones. TCO is compared with existing population-based algorithms on a set of well known numerical test functions. The experimental results show that the TCO is effective and robust; produce good results, and outperform other algorithms investigated in this consideration.