A distinctive strategy for entrainer recycling is proposed in this work for acetic acid (AA) dehydration by extractive distillation by using N methyl acetamide (NMA). The use of standard entrainers such as DMF or DMSO has the main drawback of forming an azeotrope with acetic acid. However, the vapour liquid equilibrium AA – NMA exhibits a tangential pinch point at NMA end composition. The new strategy rises from the thermodynamic analysis of the ternary diagram that which involves no azeotrope. As a result, acetic acid with high purity can be obtained by the recycling of the entrainer with a relaxed constraint in its purity. Optimization studies are discussed by using two approaches: two-step optimization method with Sequential Quadratic Programming (TSOM case) and the multi-objective genetic algorithm. The multi-objective genetic algorithm allowed the computation of the optimal acetic acid dehydration with an impurity of 3% in the recycled entrainer. Significant cost savings are achieved thanks to the optimization of both columns together. Energy consumption is reduced by 12.8% and 56.9% whereas TAC is saved by 28.4% and 56.3% compared with optimal case TSOM (impurity content 1%) and a published “Case Ref” (impurity content 0.01%), respectively.