Background RCTs are considered the reference standard in clinical research. However, surgical RCTs pose specific challenges and therefore numbers have been lower than those for randomized trials of medical interventions. In addition, surgical trials have often been associated with poor methodological quality. The objective of this study was to evaluate the evolution of quantity and quality of RCTs in pancreatic surgery and to identify evidence gaps. Methods PubMed, CENTRAL and Web of Science were searched systematically. Predefined data were extracted and organized in a database. Quantity and quality were compared for three intervals of the study period comprising more than three decades. Evidence maps were constructed to identify gaps in evidence. Results The search yielded 8210 results, of which 246 trials containing data on 26 154 patients were finally included. The number of RCTs per year increased continuously from a mean of 2·8, to 5·7 and up to 13·1 per year over the three intervals of the study. Most trials were conducted in Europe (46·3 per cent), followed by Asia (35·0 per cent) and North America (14·2 per cent). Overall, the quality of RCTs was moderate; however, with the exception of blinding, all domains of the Cochrane risk‐of‐bias tool improved significantly in the later part of the study. Evidence maps showed lack of evidence from RCTs for operations other than pancreatoduodenectomy and for specific diseases such as neuroendocrine neoplasms or intraductal papillary mucinous neoplasms. Conclusion The quantity and quality of RCTs in pancreatic surgery have increased. Evidence mapping showed gaps for specific procedures and diseases, indicating priorities for future research.