Aiming at the low conversion rate of the patent technological achievements, the paper puts forward an idea about conversion support system of patent technological achievements based on knowledge discovery in databases. Combined with Information Retrieval and Statistics, the system uses clustering, association and SVM (Support Vector Machine) to achieve the goal of the patent Data-Mining at both the macro and micro. Then, the system discovers the unknown information of the designated patent, which provides important references for the consumers such as enterprises, universities, research institutes to make decisions.