Using a time series model, we constructed a disturbance model for the aluminum sputter deposition process, and derived an autoregressive integrated moving average and recursive least-squares (ARIMA-RLS) controller based on this new disturbance model. Experimental results revealed that the ARI(3,1) model appropriately characterized the dynamic behavior of the disturbance for this process. The ARIMA-RLS controller, which includes information on process noise, is able to automatically regulate the model coefficients as the target is replaced or degrades. In this paper, the d-EWMA controller, the age-based d-EWMA controller, and the ARIMA-RLS controller were applied to aluminum sputter deposition processes in order to predict deposition rates and compare their performances. Application of the ARIMA-RLS controller is proven herein to significantly improve the estimation accuracy of the aluminum sputter deposition process, regardless of whether or not deposition rates are measured for each run.