An intelligent recognition framework combining text computing, symbolic music modeling and historical context analysis is constructed to solve the problems of low efficiency of theoretical evolution recognition and dependence on manual induction across periods in music history research. Based on theoretical literature, symbol genealogy examples and metadata, this paper establishes a multi-dimensional label database, and designs a collaborative recognition model that integrates term features, harmony-mode features and historical context features, which is used to distinguish historical periods, distinguish theoretical schools and extract evolution paths. On the historical data set of music theory, which contains theoretical texts, symbolic music examples and historical metadata, the experiment divided the training set, validation set and test set according to 8:1:1. The results show that the proposed model achieves 89.8% Accuracy, 88.9% Macro-F1 Accuracy, and 96.4% Top-3 accuracy in the music theory evolution recognition task. In the music history case verification, the average recognition accuracy reaches 90.6%, and the interpretation consistency score is 4.6. The results show that the proposed method can stably capture the linkage relationship between theoretical concepts, music example structure and historical context, and provide a computable, traceable and interpretable analysis tool for the research of music theory history.
