Abstract: The conflict, vacancy and unfeasible solutions may appear when the traditional genetic algorithms are applied into the optimizing allocation. This paper presents a novel symmetry genetic algorithm, which can avoid these defects mentioned above and succeed the fine schema from two of solutions in the father generation with two equal probabilities, thus speeding up the convergence and improving effectiveness. After being used in the examples of schematic layout in VLSI CAD, this algorithm is shown to be convergent stably and effectively.