Feature selection is an important preprocessing technique of data, which can be generally modeled as a binary optimization problem. Brain storm optimization (BSO) is a newly proposed algorithm that has not been systematically applied to feature selection problems yet. This paper studies an effective wrapper feature selection method based on BSO. Focused on this goal, firstly, a selective probability-based real encoding strategy of individual is introduced to transform the binary feature selection problem into a continuous optimization one. Based on this, then a continuous BSO-based feature selection algorithm (CBSOFS) is proposed. The proposed algorithm is tested on standard benchmark datasets and then compared to four representative algorithms. Experimental results show that CBSOFS achieves comparable results with compared algorithms.