Abstract: Chaos optimization is good at searching local solutions and sensitive to initial value. A self-adapting chaos-genetic hybrid algorithm (SA-CGA) is thus proposed to combine the chaos optimization with TGA (typical genetic algorithm) together. Differing from other conventional chaos-genetic hybrid algorithms, in the algorithm proposed a chaos operator is introduced in the evolution process according to the measure of species diversity and an excellent solvability domain is given in global search space by means of random probability. Then, as a whole, two domains are chaotically disturbed, i.e., searching in the excellent solvability domain in detail together with disturbing greatly the global solvability domain. Numerical simulation shows that the algorithm not only accelerates the convergence rate but improves its accuracy, thus solving the premature problem frequently found in TGA.