The convergence in probability of mind evolutionary algorithm (MEA) is proved by using the Markovian chain. The almost sure convergence is stronger than the convergence in probability. The migration probabilities of the similar-taxis and the dissimilation are analyzed in detail, and the evolution process of the maximal fitness function is described as a submartingale series. Based on the characteristic of the conditional expectation and the maximal fitness function, the almost sure
convergence of MEA is proved rigorously by using the sub-martingale convergence theorem.