The importance of integrating effective teaching strategies in Professional Development (PD) programs for Educational Robotics (ER)-based Science Technology Engineering and Mathematics (STEM) education is increasingly recognized. However, we need to add to the growing body of studies on comprehensive instructional approaches for teaching robotics to educators within PD STEM environments. This study investigates the effects of the modified P3 task taxonomy enriched ER-based STEM PD course on science, mathematics, and computer science teachers' STEM knowledge. The taxonomy denotes a strategy utilized to teach high school students ER. Twenty in-service teachers participated in the study, attended a 24-hour PD program in which they were taught how to use Arduino robotics kits, were assigned three tasks, and were engaged in creating lesson plans incorporating ER into their regular teaching practices. The one-group pre and post-test experimental design was adopted in the study. The instruments included a science, robotics, and mathematics content knowledge test administered before and after the PD program. Moreover, the variations in mean scores for both the pre-test and post-test, pertaining to knowledge in science, mathematics, and robotics, were demonstrated. Using a paired-sample t-test, we found that the P3 task taxonomy scaffolded PD program had statistically significant impacts with large effect sizes in robotics (2.08), science (1.49), and mathematics (0.92). These results hold important implications, suggesting that the P3 task taxonomy offers a new approach beyond learning by design, the 5E model, and project-based learning for scaffolding PD in ER settings.