Knowledge acquisition is one of the three parts of MDATA model. Knowledge acquisition mainly studies the process of human acquiring knowledge. It covers the steps of human cognition and processing information, and includes knowledge extraction and knowledge deduction. Knowledge extraction mainly studies the process of extracting entities, relationships and spatio-temporal attributes from big data in cyberspace. Knowledge deduction mainly studies the process of deducing new knowledge from known knowledge, which involves the complement of missing knowledge and the inference method of unknown knowledge in MDATA knowledge base. Firstly, in Sect. 1, the concepts of knowledge acquisition and traditional knowledge extraction and inference methods are introduced. In Sect. 2, starting from the requirements of knowledge acquisition, such as interpretability and verifiability, the challenges of traditional knowledge extraction and inference are expounded. Section 3 introduces the automatic extraction methods for MDATA models, including entity and relationship extraction and spatio-temporal attribute extraction. Section 4 introduces the inference methods for MDATA models, including the inference methods for relationships between unknown entities and temporal relationships between events. Finally, Sect. 5 summarizes this chapter.