This paper considers the control model identification and H(sub)$infty$ controller design for a tandem cold mill (TCM). In order to improve the performance of the existing automatic gauge control (AGC) system based on the Taylor linearized model of the TCM, a new mathematical model that can complement the Taylor linearized model is constructed by using the N4SID algorithm based on subspace method and the least squares algorithm based on ARX model. It is shown that the identified model had dynamic characteristics of the TCM than the existing Taylor linearized model. The H(sub)$infty$ controller is designed to have robust stability to the system parameters variation, disturbance attenuation and robust tracking capability to the set-up value of strip thickness. The H(sub)$infty$ servo problem is formulated and it is solved by using LMI (linear matrix inequality) techniques. Simulation results demonstrate the usefulness and applicability of the proposed H(sub)$infty$ controller.