The classical bacterial foraging optimization algorithm suffers from low solution accuracy and poor convergence performance due to the fixed Chemotaxis step-size. To handle these problems, an improved bacterial foraging optimization algorithm based on Levy flight is put forward. The proposed algorithm uses Levy distribution based Chemotaxis step-size for the improvement of both solution accuracy and convergence performance, and random walk strategy of Levy Flight for the improvement of bacterial migration position. Experimental results on several benchmark test functions show that the proposed algorithm achieves noticeable improvement in terms of solution quality and convergence performance in comparison with existing bacterial foraging optimization algorithms.