The invention relates to a method for controlling and optimizing the distillation tower. It controls the distillation tower in dependence on the material balance relation and component balance relation, optimizes the distillation tower by using representative model. Employing the abstraction proportion between the light product on the top of the tower and heavy product in the bottom of the tower (in following, taking the proportion between light and heavy product for abbreviation) as main controlled variable depending on the distillation principle, the material balance and component balance in the distillation tower, controlling the proportion between the light and heavy product, temperature, reflux ratio to realize stable operation by regulating the quantity of reflux and heat load of re-boiler. Simulating the practical course through the technological process simulation software, and training the nerve net model to become the representative model for the course optimization. Optimizing the operation with the representative model by using the analytical data for the raw material and product to realize the product Ka bian optimization, increasing the productivity of product with high value on the condition of guaranteeing the product quality and realizing the device energy saving and consumption reduction.