Nowadays, different types of data and information combine and interact with each other, forming a complex and huge information network. Using data mining technology, one can effectively obtain the hidden data contained in the data bureau. This technology is the most commonly used way to obtain network target data at present. In this paper, we try to practically apply related algorithms by studying the theory of multi-information fusion. Aiming at the diversity and practicality of the network, the multi-information fusion method was optimized and improved on the basis of the traditional multi-information fusion method. Secondly, a data mining system based on the concept and algorithm of association rules is established, which simplifies the working mode of frequent mining and then improves the data mining model. Finally, an empirical analysis is designed. A group of data samples are selected from the network for preliminary processing, and the data set is brought into the system for testing. From the test results, it can be seen that the algorithm designed in this paper can effectively obtain the target data and works well in a complex network environment, can analyze meaningful data association using user network rules, and provides important guidance for optimizing network information and improving extraction efficiency. This paper combines data mining technology and multi-information fusion technology to conduct in-depth research and further complete the algorithm design by combining the two technologies, which proves the accuracy and processing efficiency of the algorithm.