A firm cultural stand is the foundation of a nation, and students are undoubtedly the easiest to accept and learn about culture. In the current environment of multicultural coexistence in the world, maintaining the national culture is the bottom line of education. Therefore, this paper aimed to explore the new model of vocational education under multiculturalism based on the perspective of cultural ecological field and taking the vocational education of the IoT (Internet of Things) as the object. For the vocational education of the IoT, this paper integrated the construction of class culture and teaching design, and combined the characteristics of the vocational education and training of the IoT, so that the cultural construction runs through the whole process of vocational education. In order to better promote the development of IoT vocational education, this paper also proposed a neural network teaching evaluation method based on AHP (Analytic Hierarchy Process), which can standardize the teaching effect of teachers and evaluate the learning effect of students. The survey results showed that 97.6% of teachers and students conducted practical training in the computer room, and most of the students were dissatisfied with the current IoT vocational education model, which showed that there are still major problems in the current IoT vocational education. Compared with the traditional neural network model, the teaching evaluation model in this paper had a smaller error of 0.602-0.697. When the error was the smallest, the error of the algorithm in this paper was 17.1% less than that of the traditional neural network algorithm.