The string search problem is a classical problem of data processing. Despite a number of existing algorithms for solving this problem, the work in this direction continues. The algorithm proposed in the article develops the theoretical basis of the string search problem by combining the algorithms of two different classes, i.e. forward and backward string searching algorithms, namely, Knuth-Morris-Pratt algorithm and Bower-Moore algorithms. The paper provides the analysis of the proposed combined algorithm and comparison of its results with the basic algorithms, confirming the efficacy of the combined string search algorithm.