The invention discloses a man-machine cooperation intelligent auditing analysis method for incomplete data, and belongs to a core method in a digital auditing process. The method mainly comprises eight steps which are respectively completed by auditing personnel and a computer and sequentially comprise the following steps: constructing an auditing knowledge graph, integrating and preparing data, determining an auditing theme, generating an auditing model, generating a doubtful point list and calculating a compliance probability, confirming doubtful points and manually checking, additionally recording missing data or checking a conclusion and generating an auditing manuscript. According to the method, the limitation that the auditing cannot be carried out for incomplete data in the existing digital auditing technology is broken through, the compliance probability of the auditing point can be calculated through the current existing data, and auditing personnel are helped to determine the importance of doubtful points. After the auditor completes data additional recording, the auditing model can perform a new round of fuzzy matching and output a result, a progressive iteration type auditing point study and judgment process is formed, powerful reference is provided for the auditor to quickly identify and position problem points, and the auditor is assisted in performing compliance check.