Taking the pre-examination patents in the field of artificial intelligence in Wuxi as the research object, a patent value assessment system involving three dimensions of technology, law and market and containing 18 three level indicators is constructed. The analytic hierarchy process, entropy weight method and principal component analysis method were respectively adopted to calculate the index weights, and model comparative analysis is carried out. The results show that the entropy weight method has an accuracy rate of 95.33%, performing the best in the screening of high-value patents. It can achieve rapid screening and transformation recommendation in the pre examination stage, helping Wuxi build an intellectual property ecosystem featuring high-quality creation, high efficiency examination, and high-value transformation, and promoting the innovative development of the artificial intelligence industrial cluster.