The invention discloses an automatic modal parameter identification method for a machinery equipment structure. The method includes the following steps that a whole frequency domain section is divided into small frequency section bodies of the same width according to the section width, and statistics is conducted on the total number of poles, falling into the small frequency section bodies, of the inherent frequency; the poles, with the total number being smaller than the threshold value, in the small frequency section bodies serve as computing poles to be eliminated, and the residual poles serve as physical poles to be maintained; then, the minimum variance principle is adopted, all orders of modal parameters are identified from the remaining poles in the above steps automatically. By means of the method, the computing poles can be distinguished from the structural physical poles effectively, the modal parameters of the structure are identified automatically, the dynamic characteristic parameters of the machinery structure under the working state are obtained, manual participation in modal analysis is effectively reduced, and beneficial modes are provided for on-line detection, damage recognition and the like of the machinery structure.