In this paper, we present an ant colony optimization algorithm for load shedding minimization in smart grids. Smart grid technologies offer efficient solutions to the supply and distribution of power. However, its a great challenge to maintain the proper flow in the grid and minimize the load shedding. In reality, the situation might change dynamically and the behavior of the grid network is also stochastic. We applied our algorithm to solve the load shedding minimization problem in a deterministic setting. Our algorithm is a hybrid ant colony optimization with a constructive ant-based method followed by a local search phase. We have tested the performance of our algorithm on datasets generated using statistical data. Moreover our algorithm shows potentials for exploration in terms of the runtime and the significantly better solutions have been found.