Background: Based on the embodied cognition perspective, interdisciplinary hands-on learning combines several disciplines, such as science, technology, engineering and mathematics (STEM), to improve students' capacity to solve real-world problems. Despite the popularity of interdisciplinary hands-on learning, particularly the six-phase 6E model, more studies that include longer instructional durations, rigorous study designs and participants who understand the basic STEM knowledge and skills are required to validate this type of learning's effects on student performance. Objectives: This study developed an interdisciplinary hands-on robotics activity using a non-equivalent pretest-posttest control group design and examined its effects on high school students' learning performance. Methods: We developed 20 classes (1000 min) for a hands-on robotics activity for 80 high school students. The experimental group (EG) (n = 54) received the complete 6E model phases, whereas the control group (CG) (n = 54) received the less complete 6E model phases. All students were taught to build a robotic hand through hands-on practice before completing a robot-controlling task that involved human-computer interaction and used image recognition techniques from artificial intelligence. Results and Conclusions: The findings demonstrate that all high school students' STEM knowledge, learning motivation and creativity increased after the course. Remarkably, only if the high school students' pretest scores on STEM knowledge and learning motivation tests were lower in EG did they outperform the CG on these posttests. This finding highlights the importance of offering support to students who lack prior knowledge. The experimental group demonstrated greater creativity than did the control group. The implications of incorporating interdisciplinary hands-on learning into the 6E model's specific phases were discussed.