In this paper, a fast adaptive noise control algorithm is presented and applied to the active control of broadband noise in a one-dimensional enclosure. This algorithm employs the lattice predictor which decouples the filtered reference signals. As a result, the algorithm delivers good performance even with a wide variance in the eigenvalue spread ratio of the input correlation matrix. In addition, the algorithm can be implemented with a computational complexity that increases linearly with the filter order. To avoid the primary signal estimation process, the order-update for the estimation error is modified to obtain local estimation errors directly from the signal sensed by the error microphone. Computer simulations and laboratory experiments are conducted to investigate the convergence of this noise control algorithm in a one-dimensional enclosure. The results illustrate the fast convergence capability of the lattice-based noise control algorithm.