The invention discloses an adaptive correction method for the diagnosis experience in a rotating machinery fault diagnosis knowledge base, which comprises the following steps of collecting the fault case samples, forming the fault experience samples from the expert fault diagnosis experiences by using a symptom exhaustion method, and forming a fault sample set by using the fault case samples and the fault experience samples together; and establishing an error back propagation neural network model, and learning a function mapping relationship between the symptoms and the faults in a fault sample set through an error back propagation algorithm; solving the relation matrix between the faults and the symptoms by a learned neural network through a linear expansion method, carrying out normalization processing on the relation matrix, extracting a weight coefficient corresponding to the symptom in the diagnosis experiences and the credibility of the diagnosis experiences, thereby realizing the adaptive correction of the fault diagnosis experiences in the knowledge base. By applying the method, the diagnosis experiences can be extracted from the neural network model obtained by learning, the adaptive correction of the expert diagnosis experiences is realized, and the larger the number of the sample sets is, the higher the reliability and credibility of the obtained diagnosis experienceare.