Physical activity (PA) is increasingly recognised as an important factor within studies of osteoarthritis (OA). However, subjective methods used to assess PA are highly variable and have not been developed for use within studies of OA, which creates difficulties when comparing and interpreting PA data in OA research. The aim of this study was, therefore, to gain expert agreement on the appropriate methods to harmonise PA data among existing population cohorts to enable the investigation of the association of PA and OA. The definition of PA in an OA context and methods of harmonization were established via an international expert consensus meeting and modified Delphi exercise using a geographically diverse committee selected on the basis of individual expertise in physical activity, exercise medicine, and OA. Agreement was met for all aims of study: (1) The use of Metabolic Equivalent of Task (MET) minutes per week (MET-min/week) as a method for harmonising PA variables among cohorts; (2) The determination of methods for treating missing components of MET-min/week calculation; a value will be produced from comparable activities within a representative cohort; (3) Exclusion of the domain of occupation from total MET-min/week; (4) The need for a specific measure of joint loading of an activity in addition to intensity and time, in studies of diseases, such as OA. This study has developed a systematic method to classify and harmonise PA in existing OA cohorts. It also provides minimum requirements for future studies intending to include subjective PA measures.