We investigated the 12-month intra-person variability of various physical activity assessments used in the Kuopio Ischaemic Heart Disease Risk Factor Study (KIHD) in 51 men aged 54 years. The methods were a 24-hour total activity recording, a 12-month leisure time activity history, a 7-day leisure time activity recall and a habitual occupational activity interview. More leisure time activity was reported in the retest. Intra-person variability was high for all physical activity indices. The proportion of the mean absolute test-retest difference out of the test-retest mean was 18% for the 24-hour total activity recording, 45% for the 12-month leisure time activity history, 86% for the 7-day leisure time activity recall, 70% for kilometres of conditioning activity per week (No. = 51) and 22% for the occupational activity interview (No. = 39). However, respective intraclass correlations for these indices were 0.43, 0.58, 0.35, 0.71 and 0.69. The physical activity assessments differ in observed intra-person variability in accordance with the type of activity, time frame and the method of recall. Owing to its representative time frame and relatively small intra-person variability the 12-month history may be recommended for a standard instrument for population studies concerning leisure time activity.