The essence of simulated annealing algorithm is getting away from the local optimum and converging to the global optimum,so the method is introduced into genetic algorithms.Combining simulated annealing and genetic algorithm together,a new algorithm called adaptive genetic simulated annealing algorithm is formed and is introduced in order to solve inverse problem related to ellipsometric measurement of thin films.The hybrid algorithm combines two powers:global optimization and local search.What is more,in order to check the feasibility and efficiency of the algorithm in ellipsometric measurement of thin films,numerical simulation is done in this paper.The results obtained may be a good example to demonstrate the feasibility and efficiency of the approach to solve more complex inverse problems in ellipsometry.