Controllability of a nonlinear high-purity distillation column using nonlinear control algorithms is addressed in this study. A modeling and control approach is studied in this paper to handle both nonlinearity and interactions of the column. A low-order empirical nonlinear model is established and utilized to construct nonlinear simplified decouplers, thereby diagonalizing the input-output mapping of the column model. Linearizing feedback transformations are then adopted to reduce process nonlinearity for the nearly diagonalized system. Finally, PI controllers are designed on the basis of this transformed linear system, Significant performance improvement in control tests verifies the effectiveness of the proposed nonlinear control approach.