Experimental modal analysis (EMA) is a program that allows structure modal parameters to be extracted from the measured response under impulse excitation. Traditional EMA method is based on the assumption of linear mode theory, which is easy to cause pseudo-mode. Therefore, it is necessary to propose a de-theorizing EMA method. In order to solve the problem of obtaining modal parameters without theorizing, a method using data mining is proposed in this paper. As a data mining method, symbolic regression can obtain modal parameters by only mining function expression rules from single point response data. Taking modal parameter extraction based on modal theory as a reference, we designed a simulation test and an experiment to verify the feasibility of the proposed method. In addition, it has a good application prospect for de-theorizing modal rule mining based on data.