In the complex information environment, intelligence data has more sources, more complex structure and larger scale. How to identify information with intelligence value from multi-source data in time is an urgent problem to be solved. This paper deeply analyzes the new requirements of multi-source data intelligence value discovery in complex information environment. On this basis, a multi-source data information value discovery model is proposed. Firstly, a dynamic multi-source data space is constructed. Secondly, a design scheme of information value discovery is proposed from three dimensions: external basic characteristics of multi-source data, content semantic characteristics and user information demand characteristics. In complex information environment, the intelligence value discovery of multi-source data needs fast perception, fine-grained mining, fine judgment and intelligent analysis. It is necessary to always pay attention to the intelligence needs of users, and combine deep learning and other technologies to achieve the efficient, accurate and intelligent development of multi-source data intelligence value discovery.