The application discloses an intelligent teaching method fusing sound, meaning and object, and the method is characterized in that: initial course sets are matched by quantized scoring and ability grading matching through acquisition of evaluation
questionnaire data, teaching path sequences are generated through a path planning
algorithm, mapping is performed from five sensory
perception dimensions and structured courseware sequences are assembled, real operation videos are received after the courseware is presented to a parent terminal, test result reports are generated through multi-
modal three-level
cascade evaluation, and it is judged whether to trigger path adjustment and update the ability vector to regenerate the adjusted path according to the
test result report. The application solves the technical problems of single teaching dimension, weak generalization ability and lack of personalized
adaptation in the prior art, and meets the strict quality requirements of family intervention for children with spectrum disorders.