Science and technology big data contain a lot of cross-media information.There are images and texts in the scientific paper.The s ingle modal search method cannot well meet the needs of scientific researchers.This paper proposes a cross-media scientific research achievements retrieval method based on deep language model (CARDL).It achieves a unified cross-media semantic representation by learning the semantic association between different modal data, and is applied to the generation of text semantic vector of scientific research achievements, and then cross-media retrieval is realized through semantic similarity matching between different modal data.Experimental results show that the proposed CARDL method achieves better cross-modal retrieval performance than existing methods. Key words science and technology big data ; cross-media retrieval; cross-media semantic association learning; deep language model; semantic similarity