BACKGROUND: Multiple sclerosis is a common inflammatory neurological disease among young adults and is the tenth leading cause of the global burden of disease. Existing common treatments such as pharmacological and palliative therapies do not control the neurodegenerative process or cure multiple sclerosis. Numerous epidemiological surveys, randomised controlled trials, and systematic reviews with meta-analyses support the effects of physical activity on health-related outcomes among patients with multiple sclerosis. Moreover, bibliometric analysis can provide a broad evidence synthesis beyond systematic reviews and meta-analyses, allowing researchers and other stakeholders to obtain a one-stop overview of this research field. Therefore, this bibliometric analysis aims to provide insight into the knowledge structure of the field of physical activity for multiple sclerosis over the past three decades, and to predict emerging research trends. METHODS: This study strictly complied with step-by-step guidelines of bibliometric analysis, combining performance analysis and science mapping. Four indexes from the Web of Science Core Collection were selected as data sources, and articles and review articles in the field of physical activity for multiple sclerosis from 1994 to 2023 were included in this analysis. Mircrosoft Excel, RStudio, VOSviewer 1.6.20, and CiteSpace 6.3.R1 (64-bit) Advanced were used to perform performance analysis and science mapping. RESULTS: Over the past three decades, this field published a total of 1,271 documents, with the scientific output showing a rapid upward trend over the past two decades. Robbert W Motl was the most prolific author in this field, with a total of 300 publications. The USA contributed nearly half of the publications in this field (549 documents), and the University of Illinois System was the institution with the highest number of publications (222 documents). Multiple Sclerosis and Related Disorders was the journal that published the highest number of documents in this field (117 documents), while more than a third of this field's publications were included in the category: Clinical Neurology (438 documents). The Reference co-citation analysis identified three main research trends, including shifts in research methodology, changes in health outcomes in randomised controlled trials, and shifts in different types of physical activity interventions. Combining the results from reference co-citation analysis and citation burst analysis, the combination of behaviour change technique and telerehabilitation may be the emerging research trend. CONCLUSION: This bibliometric analysis identifies rapid growth in the field of physical activity for multiple sclerosis over the past two decades. Moreover, the combination of performance analysis and science mapping provides insight into knowledge structure in this field and informed future research trends for researchers and the relevant stakeholders.