Information technology’s continuous development and progress can bring great changes to society, which is also born in the hierarchical protection of information systems. Among them, grade protection evaluation is the key part, and methods to find useful information from its complicated evaluation data to assist decision-making have become an important content of research. Aiming at the evaluation process of data mining, this article intends to analyze the classical algorithms of association rules. The Apriori algorithm and FP-growth, two common and representative algorithms, are chosen as a comparison. Finally, it is found that the performance of the latter is better than that of the former. The rationality of association rules is proved by analysis to better assist the evaluators in the evaluation work and reduce the misjudgment of the evaluation results.