A multilevel hierarchical control system has been designed and is being applied to chemical-mechanical planarization (CMP) process control, The current implementation of the control system incorporates closed-loop run-to-run (R2R) control and open-loop real-time monitoring, and can accommodate inter-cell control. The R2R control element is enabled via a generic cell controller (GCC) implementation that provides flexible automated control of the process and equipment, multiple control algorithm branches and fuzzy logic decision capability among the branches, simulation capabilities, hardware and software independence, and extensive GUI support for control and data analysis. The R2R element utilizes a linear approximation multivariate control algorithm (branch) that supports individual exponential weighted moving average (EWMA) modeling of advices (outputs), weighting of inputs, granularity, and input bounding. The real-time element of the control system utilizes a partial least squares (PLS) algorithm to identify real-time equipment input trace patterns and relate these patterns to alarming conditions. The entire control system is designed to provide multivariate control of CMP process removal rate and uniformity. As a result of extensive design of experiments and testing, the R2R control level has been demonstrated to achieve good control of removal rate and fair control of uniformity. The addition of the real-time element is expected to improve process control and reduce R2R process noise, thus leading to a more effective R2R control element.