Enlightened by the conception of ethnic group in social science and a perspective to analyze the structure and evolutionary tendency of population in terms of ethnic group, a population structured technology, ethnic group mechanism, is proposed. Meanwhile, an ethnic group evolution algorithm (EGEA) with a dual track co-evolution process and special ethnic group operators is designed for binary coding. The simulation tests of the classical function and challenging composition test function show that the EGEA can restrain premature convergence effectively during the evolutionary process while improving the search efficiency greatly. The comparisons between EGEA and other typical algorithm show EGEA is a competent algorithm for solving numerical optimization problems.