The experimental evaluation and design model of C4 olefin preparation based on high dimensional discrete space optimization was established in this paper. First, the relationship between ethanol conversion, C4 olefin selectivity and temperature was studied by fitting. Then, we took catalyst combination and temperature as independent variables, ethanol conversion and C4 olefin selectivity as dependent variables, conducted partial correlation analysis on the latter to observe their correlation degree, further used multivariate analysis of variance, and conducted inter-agent effect test. The effects of catalyst combination and temperature on ethanol conversion and C4 olefin selectivity were evaluated by the sum of type III squares. Subsequently, we used the control value of the aforementioned dependent variable to generate a high-dimensional discrete space, conducted stepwise regression of the aforementioned dependent variable and independent variable, and then tested its accuracy according to the objective function of maximizing the yield of C4 olefin. Combining with the constraints of discrete variables, the method of ergodic and genetic algorithm is used to search for the optimal value in high dimensional discrete space, and the corresponding results are obtained. Finally, as an innovative supplement, we used Pareto optimization to explore the method of producing more C4 olefin by ethanol coupling experiment under the same amount of ethanol substance as possible.