Simulated annealing is a new kind of random search methods developed in recent years.it canalso be considered as an extension to the classical hill-climbing method in AI——probabilistichill-cimbing.One of its most important features is its global convergence.The convergence ofsimulated annealing algorithm is determined by state generating probability,state acceptingprobability,and temperature decreasing rate.This paper gives a generalized simulated annealingalgorithm with dynamic generating and accepting probabilities.The paper also shows that thegenerating and accepting probabilities can adopt many different kinds of distributions while the globalconvergence is guaranteed. Simulated annealing is a new kind of random search methods developed in recent years. It canalso be considered as an extension to the classical hill-climbing method in AI-- probabilistichill-cimbing. One of the most important features is its global convergence. Convergence ofsimulated annealing algorithm is determined by state generating probability, state acceptingprobability, and temperature decreasing rate. This paper gives a generalized simulated annealing algorithm with dynamic generating and accepting probabilities. The paper also shows that thegeneratingand accepting probabilitiescantake many different kinds of distributions while the globalconvergence is guaranteed.