This study aims to formulate an effective educational plan by conducting an in-depth analysis of elementary school teachers’ perceptions and demands for convergence education using physical computing and artificial intelligence (AI) in elementary science education. Toward this end, it conducted a survey and interviews on with a group of 36 expert teachers who participated in Digital Sprout Camp, a group of 36 randomly selected general teachers, and a total of 72 elementary school teachers. The results demonstrated that all teachers were positively aware of the necessity and effectiveness of convergence education. In particular, the expert group provided high evaluations on the benefit of convergence education, and the execution rate was high. However, the study confirmed a low frequency of execution for the general teacher group, which was a comparative group, due to limitations such as lack of expertise, experience, training, environment/teaching aids, and preparation time in the actual class application process. The teachers also recognized that convergence education in physical computing and AI significantly enhanced students’ scientific inquiry ability, creativity, problem solving ability, and learning initiative, effectively improved the acquisition of scientific concepts, and induced interest. However, the teachers pointed out the high initial cost of introduction, issues in maintenance, and lack of access to teaching aids as major obstacles. In addition, they mentioned the necessity of providing systematic and practical training in convergence education, developing teaching aids and materials suitable for elementary schools, and distributing class materials that can be used immediately. In addition, they emphasized the need to systematically improve the curriculum and introduce an evaluation method (e.g., portfolio) that reflect this to foster the connection between convergence education and the objectives of science education. The study deems their responses meaningful in establishing the basis of the future development of policies and programs for convergence education in elementary schools, as they shed light on the actual application status and limitations of physical computing and AI-based convergence education.