Industrial Internet and intelligent manufacturing put forward higher requirements for talents while promoting industrial upgrading, and there is a problem of insufficient number of automation technology talents in the field of industrial Internet. This study completes the education reform of the electrical automation technology specialty from two aspects: the innovation of the curriculum system and the docking between the specialty and the industry. The K-means clustering algorithm is used to cluster the students’ learning data to find out the problems of students’ learning in the electrical automation technology specialty, which provides a scientific basis for the education reform. Logistic regression model was also used to analyze students’ satisfaction with the reform of electrical automation technology. The results of the study show that the learning effect of students can be improved by increasing the number of hours of after-school practical learning and constructing the learning mode of bringing good to bad. Training base selection, practical training process, internship process and intelligent manufacturing facilities all positively affect students’ satisfaction with the results of teaching reform, with training base selection having the greatest influence and intelligent manufacturing facilities having the smallest influence, with a coefficient of 0.735. It indicates that the teaching reform strategy in this paper is positively successful overall, and provides valuable decision-making information for the teaching reform of electrical automation technology.