The objective of this study was to assess the between-day repeatability of knee kinematics during activities of daily living recorded by electrogoniometry. One rater assessed the peak knee angles and knee excursion of 15 subjects during 13 activities twice with an average of 22 days (range 5–31) between the two assessments. The 15 subjects included four patients one year after total knee replacement (TKR) surgery, five patients before TKR surgery and six age-matched controls. Intra-class correlation coefficients and Bland and Altman coefficient of repeatability were derived to analyse the results. Only the most affected leg of the patients and the right leg of the controls were used for analysis. Different measures of repeatability showed different results. Intra-class correlation coefficients were higher than 0.75 for peak values of all functions except sitting down and rising from a standard chair. However, coefficients of repeatability ranged from 5.6° for the loading response in level walking to 39.8° for stepping out of a bath. Both of these values are higher than clinically significant changes seen after total knee surgery. It was concluded that for a single assessment on individual patients, the functional knee motion as performed in this study did not have sufficient repeatability. However, if the measurements are used to assess the average changes before and after surgery in a group of patients, the assessment of knee motion during activities such as level walking, and slope and stair ascending and descending were found to be sufficiently repeatable.