This work explores the use of variable measurement and control structures in the state estimation and control of distillation processes. Distillation processes with complex behavior are difficult to control (azeotropic and extractive separation) due to the existence of sharp temperature gradients and were found to be ideal candidates for the implementation of variable control structures. An on-line sequential analysis of the temperature measurements along the length of the column allows a fast selection of the best measurements to be used in low order, simple and robust control schemes. Overall results indicate that variations in column operating conditions could also cause variations in the resulting control schemes. For the case of extractive distillation, more robust estimation and control were obtained through the use of the proposed methodology.