Abstract: Online handwritten mathematical expression recognition method faces some challenges,such as the uncertainty of writing characters,the complexity of mathematical expression structure and the various written styles of expression.Existing no-human-involved recognition algorithms have low recognition accuracy,espe-cially when accidental errors and complex structures occur in mathematical expression writing.A human-in-the-loop based online handwritten mathematical expression recognition method is proposed,which mainly intro-duces human participation in the structural analysis stage to define and perfect the structural strokes or strokes within the mathematical structure,by means of human modification for the ambiguous strokes in the structure and operation of structural stroke makeup.To evaluate the effectiveness of the proposed method,the method is com-pared with a no-human-involved recognition algorithm in rate of structural analysis and expression recognition rate.The results show that proposed method can effectively promote users to participate in the process of recogni-tion,and that incorporating user in the recognition processes can improve the accuracy of recognition of structure(9.26%)and expression(13.99%)in handwritten mathematical expression.