The invention discloses a metering acquisition research direction trend analysis method based on natural language processing, and relates to the field of natural language processing. At present, the traditional manual query, reading, analysis and the like are difficult to accurately analyze the development trend of a new technology. The method comprises the following steps of (1) designing a crawler program, and obtaining the core journal papers in the field of measurement and collection; (2) carrying out word segmentation processing on the obtained paper text content by adopting a statisticallanguage model; (3) filtering out the stop words through the part-of-speech tagging; (4) in combination with the TF-IDF and TextRank methods, extracting the keywords from a candidate word bank; (5) judging the correlation between the keywords, and combining different vocabularies for expressing the same object; and (6) drawing an evolution dynamic graph of the corresponding research hotspot basedon the extraction and combination results. According to the technical scheme, the increasingly huge literature database resources at present can be better adapted, and the research hotspots in the related research directions of measurement and collection can be rapidly and intelligently extracted from the literature database resources.