Spatial perspective-taking (SPT) ability positively influences performance in STEM fields. While limited research studies have been done with school students, they have yielded inconclusive findings and, hence, here we report findings from our study with an augmented reality (AR) enhanced learning environment (ARELE), GeoSolvAR, on middle-school students' SPT ability. We conducted a sequential explanatory mixed-method study with 90 students to test GeoSolvAR's twin features of augmented systems and inquiry-based activities. The study had one experimental (n = 30) and two control groups (n = 30 × 2), where students were individually administered an online pre- and posttest to measure their SPT ability. The experimental group worked with both the features of GeoSolvAR, while the control group worked with only the AR feature or had no intervention. Our findings reveal that the experimental group showed significant improvement in SPT ability while students in the control groups showed non-significant improvement or no change. Furthermore, we found that all experimental group students used certain body movements and imagined shifts in spatial positions while solving the activity questions. Thus, we conjecture that GeoSolvAR affords thinking and imagining with the body which makes it a potent tool for improving students' SPT ability.