In this study, a quantum-behaved particle swarm optimization (QPSO) based on hybrid evolution (HEQPSO) approach is proposed to estimate parameters of chaotic dynamic systems, in which the proposed HEQPSO algorithm combines the conceptions of genetic algorithm (GA) and adaptive annealing learning algorithm with the QPSO algorithm. That is, the mutation strategy in GA is used for conquering premature; adaptive decaying learning similar to simulated annealing (SA) is adopted for overcoming stagnation problem in searching optimal solutions. Three examples are illustrated to estimate parameters of chaotic dynamical systems using the proposed HEQPSO approach. From the numerical simulations and comparisons with other extant evolutionary methods in Lorenz system, the validity and superiority of the HEQPSO approach are verified. In addition, the effectiveness and robustness of parameter estimations for Chen and Rossler systems are demonstrated by the proposed HEQPSO approach.