The work aims to systematically investigate the user requirements for programming education Apps of three key groups—primary school students, parents, and teachers, and propose interface interaction design strategies to enhance the effectiveness of programming education Apps, so as to address the issues of poor usability and suboptimal user experience in programming education Apps for primary school students. An initial feature list for programming education Apps was developed based on the journey map and the digital education theory LM-GM model. A dual-factor Likert scale was created using the Kano model analysis method, and a survey was conducted with primary school students, parents, and teachers. The SI and DSI indices for each product feature were calculated for the three user groups, resulting in the formation of a Better-Worse four-quadrant model. The study identified five essential attributes, eight expected attributes, and twelve attractive attributes for programming education Apps. The design and development of the Apps WannaBone1 and WannaBone2, which catered to the needs of three major user groups, included 40 learning tasks, 13 interface modules, and two task interface design modes, integrated AIGC with smart educational technology, highlighted multi-sensory stimulation, visual programming interaction, and a driven motivation mechanism. Designing programming education Apps based on the Kano model analysis method allows for a deep understanding of the multidimensional psychological needs of the target user groups. This approach enables the creation of interface interaction designs that offer an excellent user experience and provides innovative design ideas and practical guidance for primary school students' artificial intelligence education.