Background Panoramic video as a novel instructional media with increasing educational application. However, empirical research on panoramic video instruction is mostly episodic with conditional and inconclusive results. Objectives This study reviewed 10 years (2011-2021) of research on panoramic video in education to synthesize, meta-analyse, and critically evaluate the state-of-art research findings, which can inform future educational practice and research directions. Methods Following the protocol of systematic review, this study selected 83 articles from four major research databases and analysed them from five interrelated aspects: instructional context, technological features, instructional design, research design, and instructional effect. A meta-analysis was also conducted to determine the overall effectiveness of panoramic video on education and the moderating factors. Results and Conclusions The results revealed a rapid upward trend in publication since 2018 with unequal distribution among regions and learning domains. Despite the technical advancement in immersion and interaction, panoramic videos were found to mainly support the acquisition of content and factual knowledge through direct instruction, and lack in-video scaffolding and assessment. Moreover, many interventions were one-time, short-term studies that are susceptible of novelty effect. Lastly, an overall medium effect (g = 0.711) of panoramic video was reported and six significant moderating factors were identified based on the meta-analysis results. Major takeaways and implications Panoramic videos are ubiquitously effective across grade levels, learning domains, and pedagogies, but are more suitable for teaching conceptual knowledge. Design features such as interaction, scaffolding, assessment, iteration, and video length deserve our special attention in panoramic video instruction.