Summary: Aiming at a disadvantage of traditional differential evolution, namely, existence of stagnating individuals in the evolutionary process, a novel differential evolution algorithm based on elite-cloud mutation is proposed in this study. In the proposed algorithm, stagnation generation of each individual is counted in the evolutionary process. Moreover, an individual is executed by elite-cloud mutation to approach the best individual when the stagnation generation of the individual is more than a pre-defined threshold value. Thus, it can accelerate the convergence speed. Additionally, in order to increase the population diversity, it executes the opposition-based learning operator with a certain probability. Experimental results indicate that the proposed algorithm obtains promising performance in both solution precision and convergence speed.