Background: Valid measures to assess either small or assisted performed movements of people with profound intellectual and multiple disabilities (PIMD) are required. We analysed the construct validity of the Actiwatch-2 to assess movement in people with PIMD. Method: Twenty-two persons with PIMD were video recorded while wearing an Actiwatch-2. We used 15s-partial-interval recording to record upper body movement, body position and activity situation. Multilevel analyses were used to evaluate if the Actiwatch-2, based on produced counts, could detect changes in these factors. Results: The presence versus absence of upper body movement and an activity situation in which participants were involved versus not involved resulted in significantly higher counts, with a large variety in predicted counts between participants. No relationship between body position and counts was found. Conclusions: The Actiwatch-2 seems able to assess obvious upper body movement in people with PIMD, and whether there is involvement in an activity situation.