The dynamic scheduling (DS) strategy based decoding algorithm for low-density parity-check (LDPC) codes is explored from a novel point of view. In the view of dynamic selection strategies, intelligent search algorithms are taken into consideration. As is known, the dynamic decoding algorithm is a greedy one, where the message with the maximal-residual is always propagated, which causes unfairness in decoding scheduling when updating messages. To cope with the problem, the simulated annealing (SA) method is first introduced to provide assistance in dynamic scheduling-based decoding. Then, a more accurate dynamic selection strategy is presented based on the SA-assisted dynamic scheduling strategy, SA-DS. When the Metroplis acceptance criterion is satisfied, the SA-DS algorithm can select the most unreliable node accurately and give a priority to update it. Simulation results demonstrate that the proposed algorithm behaves better than any other algorithm of interests.