Abstract: In order to plan the industrial structures in the three northeastern provinces, to increase the three northeast provinces regional economic industry's overall competitiveness, the industrial structures in the three northeastern provinces are studied by the author using Markov prediction method. And the interval adaptive genetic algorithm is employed to solve the Markov state transfer probability matrix. According to 2004-2014 statistical data of industrial structure, the first, second and third industrial structures in the three northeastern provinces are predicted. The average error is 2.92 percent. The industrial structures in the three northeastern provinces in 2015-2016 are predicted using this model, it provides a certain amount of theoretical basis for making corresponding local industrial structure policies in the three northeastern provinces.