Abstract: In order to solve the route search problem in stochastic Petri net,the feature of reliability is introduced into each transition,and suppose that the lifetime distribution of a transition obeys the Weibull distribution model,then a memorial stochastic Petri net with reliability whose elements can record a little information is proposed based on the ant colony optimization algorithm and features of the memorial stochastic Petri net.When the Petri net is running,enough tokens walk and leave pheromone in the Petri net,which is used to adjust the route choice of tokens.When certain reliability requirement needs to be satisfied,ants will leave more pheromone on routes with less delay and required reliability,at last a clear ant route can be found on the route with the least delay and required reliability.Because the reliability of a transition will decrease with time going on,when the optimum route can’t satisfy certain reliability after a while,it needs to choose another new optimum route that satisfies certain reliability and has the relative least delay.The result of simulation shows that tokens can effectively find the best ant route in a timed stochastic Petri net,and they can also search the new optimum route as the change of transitions’ reliability.