The clustering analysis method is used to transform the dynamic comprehensive evaluation of multiple experts into the static comprehensive evaluation problems. The scatter degree method is adopted to determine the weight of each index objectively, combined with the subjective weight, and the mathematical programming method is used for the index integration. Finally, Bayesian networks model is used to make a classification evaluation of 24 scientific and technological achievements, and give the probability measurement of each scientific and technological achievements awarded to a certain place and rank within each level. The empirical research suggests that most of Chinese scientific researches have achieved the fruitful results with high research values while lack of creativity and breakthroughs.