Traditional　method　about　designing　blue　print　in　course　of　surface　laser　quenching　was　studied.　Expert　system　of　laser　quenching　based　on　genetic-neural　network　was　presented.　Knowledge　acquisition,　construction　of　knowledge　base　and　design　of　inference　engine　were　introduced　in　detail.　The　trial　running　results　show　that　this　research　raised　an　advanced　and　reasonable　path　for　performance　prediction　and　optimization　design　of　laser　quenching.