The values used for natural mortality (M) are very influential in stock assessment models, affecting model outcomes and management advice. Natural mortality is one of the most difficult demographic parameters to estimate, and there is often limited information about the true levels. Here, we summarise the evidence used to estimate natural mortality at age for the four main stocks of yellowfin tuna (Indian, Western and Central Pacific, Eastern Pacific, and Atlantic Oceans), and identify important issues and information gaps. We describe the history of natural mortality values used in stock assessments by the tuna Regional Fisheries Management Organisations responsible for managing each stock and assess the evidence supporting these values. In June 2021, an online meeting was held by the Center for the Advancement of Population Assessment Methodology (CAPAM), to provide advice and guidance on practices for modelling natural mortality in fishery assessments. Based on approaches presented and discussed at the meeting, we develop a range of yellowfin tuna natural mortality prior distributions for each stock. We also recommend future research to improve these estimates of natural mortality.