The frequent switching of tooling/fixture， tool/gauge and system utilizationse cases in the multi-variety mixed-model production line under the mode of "industrial 4.0" intelligent manufacturing can easily lead to abnormal production and short-term stopping risk in workstations. First， the mean residence time（MRT） was proposed by analyzing the uncertainties caused by product polymorphism in the multi-variety mixed-model production line. Next， this study by considerings the minimum production cycle of the MRT multi-variety mixed-model production line， establishes the product operation schedule based on the sequence of products entering the production line， and establishes an intelligent sequencing mathematical model by introducing the key variables of MRT to the objective of minimal cycle optimization. After that， based on the Palmer principle， an improved genetic algorithm was designed to solve the intelligent sequencing problem of the multi product mixed-model production line. Finally， the feasibility and validity of this method are verified by an example.