Abstract: To separate source signals one by one, we consider the blind source extraction problem of instantaneous mixtures using artificial bee colony optimization algorithm. A goal function is constructed first by exploiting the higher statistics prosperities of source signals, then the optimal extracted vectors are determined through maximizing the goal function using artificial bee colony optimization algorithm, so as to separate source one by one. The simulation results show that the method can achieve the blind separation for mixed signals in decreasing order of absolute kurtosis. Moreover, it can separate images of sub-gaussian distribution and speeches of super-gaussian distribution.