The invention discloses a power transmission and transformation equipment defect data machine self-clustering tool based on a machine learning algorithm. The self-clustering tool consists of a data resolver (1), a clustering device (2) and a result processor (3) for processing power transmission and transformation equipment defect data according to different classifications. The self-clustering tool has the capacity of classifying the conventional mussy power transmission and transformation equipment defect data, can extract key information from character descriptions and is self-learning and self-improved. The main core algorithm is that Chinese word identification (MMSEG), a term frequency inverted document frequency algorithm (TFIDF) and a supporting vector machine (SVM) in two variable bodies of the algorithm are matched with one another to the maximum extent; MMSEG is used for resolving knowledge; the TFIDF is used for estimating the importance of a certain type of information in information; and SVM is used for processing machine learning problems.