Lameness is common in sows, causing production losses and reduced welfare. We studied how lameness affects behaviour in gestating sows and how treatment with oral ketoprofen affects the behaviour of lame sows. Sows were group-housed in partly slatted pens. On day 0, sow lameness was clinically evaluated using a scale from 0 to 4 (no to severe lameness). Sows with non-infectious lameness (L, score 2-3, n = 13) and control sows (C, score 0, n = 13), 1-3 L:C pairs per pen, were followed in their home pens. L sows were given a placebo (P, n = 4), or ketoprofen (K) 2 mg/kg (n = 5) or 4 mg/kg (n = 4) orally once a day for 5 days. Behaviour was observed for 2 h by direct scan sampling on days 0 and 5, starting 10 min after feeding (2:40 pm). Data were analysed with Mann-Whitney U-tests. The two K levels were pooled as there was no difference in their clinical effect. L sows lay down more (p = 0.02) and stood (p = 0.05) and explored pen features (p = 0.005) less than C sows on day 0. The behaviour of L sows showing an improvement in their lameness score (n = 7, 1 P sow and 6 K sows) after the treatment did not differ from C sows on day 5 (p > 0.05). However, sows showing no improvement (n = 6, 3 P sows and 3 K sows) laid down more (p = 0.01), while moving (p = 0.02) and standing (p = 0.04) less than C sows. Lameness reduced activity in group-housed gestating sows, indicating discomfort or pain. Improvement in the lameness score is connected to normalisation of the behaviour. As ketoprofen improves the success of recovery, treatment of lame animals is recommended.