This paper presents a hybrid algorithm which based on genetic algorithm and tabu search algorithm to optimize the phasor measurement units(PMU) placement problem. Genetic-Tabu Search combines the random searching ability and parallel of Genetic Algorithm with the memory function of Tabu Algorithm, and solves problems of Genetic Algorithm such as bad grade ability, Pre-maturity effectively, and enhances the convergence speed and utilization quality. Meanwhile, the category operations of Generic Algorithm keep the advantages of multi-origin of Generic Algorithm and amend the weak point of Tabu Searching which is the lack of parallel for single-single operation. It takes the infeasible and heuristic restoration while the constraint is handling. The Genetic-Tabu Search algorithm is applied to PMU placement optimization and fulfills the requirement of minimizing the number of PMUs in the system while the all node voltage phasor observable. A graph-theoretic procedure based on depth first search is adopted to analyze the system observability. Test results show that the GATS algorithm is effective.