An adaptive damping control scheme based on classification and regression tree (CART) using wide-area signals is proposed in this paper. Different polytopes are chosen using classification trees. Next, a family of robust polytopic controllers is designed off-line. Finally, using regression trees and PMU data, the desired polytopic controller is identified in real time. A 16-generator, 68-bus system is used as the test system. Remote signals obtained from PMUs are employed for control purposes. The simulation results demonstrate that the proposed adaptive control scheme is able to provide adequate damping for the oscillation modes of interest with respect to varying operating conditions without any prior knowledge of the post-disturbance state.