Abstract: Aiming at the defects formed in the manufacturing process of additive, it will cause irreversible influence on the workpiece. The current signal and voltage signal of CMT additive manufacturing process were analyzed, and an online monitoring method of CMT additive manufacturing defects based on time series algorithm was proposed. Different welding conditions were set, the original current and voltage signals of good group and defective group were collected, and SAX (Symbolic aggregate approximation) algorithm was used to preprocess the data. Random forest model was used to reclassify numerical data to achieve real-time monitoring effect. At the same time, in order to highlight the superiority of SAX algorithm, a comparative test group was set up, and the original current data was directly put into the random forest model for classification. The experimental results showed that the accuracy of the test set of the original current group was 80%, and that of the SAX algorithm data preprocessing group was 96%.