With the rapid development of virtual technologies, there is a growing body of literature investigating the impact of virtual technologies on students’ spatial ability. However, it remains unclear whether virtual technologies can effectively improve students’ spatial ability. Therefore, this meta-analysis was conducted to synthesize the findings on the overall effects of virtual-based spatial ability enhancement. We systematically searched literature published from 2010 to 2020 (excluding non-empirical articles) and found 36 experimental peer-reviewed journal articles that met the inclusion criteria. Then, the random-effects model (REM) was used to calculate the pooled effect size. Results showed that virtual technologies have a medium effect on developing spatial ability with an overall effect size of 0.617. The studies were also coded to examine the moderating effects of their characteristics, such as learner stage, virtual technologies, disciplines, experimental design, learning application types, spatial ability, and testing instruments, on the outcome measure. The moderator analysis indicated that the virtual-based spatial ability improvement was more effective (a) for preschool learners, (b) in the fields of natural science and engineering technologies, (c) for all types of spatial ability, and (d) when learning during 3 to 6 months. Furthermore, augmented reality was most conducive to improving learners’ spatial ability compared with other virtual technologies. These findings provided insights for future studies and practices on using virtual technologies to cultivate spatial ability.