“Numerical Intelligence” teaching is an important content and means of classroom teaching reform in clothing design and craftsmanship. The first part of the study focuses on the use of digital intelligence technology in trouser design. After selecting the characteristic parameters of the trouser structure, the GA algorithm is used to improve the GRNN algorithm, the global optimization of the smooth factor is carried out, the trouser structure generating a model based on GA-GRNN is constructed, and the model is trained and tested in order to analyze its performance. With this background, we explore the combination of numerical intelligence technology and clothing craft course, innovatively propose the teaching mode of numerical intelligence craft course, and design teaching experiments to explore the practical effect of this teaching mode. The GA-optimized GRNN model improves the number of iterations and the adaptability by 63.23% and 0.0131, respectively, and the error between the model test generation results and the actual values is below 0.4, which has a better trouser suit parameter size generation effect. After applying the teaching mode of the digital intellectualization process course, the student’s overall evaluation of the mode is better, and the number of students with good and excellent academic performance has increased by 64.29%. The teaching mode of the digital-intelligent process course can promote students’ learning efficiency and improve the quality of teaching in the process course.