Loose housing farrowing environments for sows are common in conventional production in most Scandinavian countries and in organic production across Europe in order to fulfil the sows´ behavioural needs. However, the most common cause of piglet mortality in loose housed farrowing environments is crushing by the sow. This study compared the relationship between sow behaviour and piglet mortality for loose housed sows in two farrowing environments: outdoors in huts on pasture and indoors in farrowing pens. Forty Yorkshire x Swedish Landrace sows were videotaped inside the hut (3.9m2, parity 1 and 3) or in the pen (8.2m2, parity 2 and 4) from 12 hours before farrowing until 2 hours after the last piglet was born. The statistical analyses included LSM and regression analysis. Farrowing duration was shorter (3.5 vs. 5.2 hours, p<0.001, N=90) and the probability of stillborn piglets was lower (0.11 vs. 0.34, p<0.01, N=132) when sows farrowed in huts compared with pens. The regression coefficient for piglet mortality (until day 4 after birth) on sow behaviour during farrowing differed between farrowing environments (percent active: b=2.26 vs. -0.45, p<0.001; number of posture changes: b=1.46 vs. -0.82, p<0.01 in huts and pens respectively) indicating that restless sow behaviour leads to high piglet mortality when sows farrowed in huts, but to low piglet mortality when sows farrowed in pens. There was no significant effect of parity (nested within environment) for the investigated parameters. We conclude that parturition progress is more rapid and stillbirth frequency lower when sows farrow outdoors in huts compared to when they farrow indoors in pens. Moreover, restless sow behaviour (long activity duration and high frequency of posture changes) during farrowing is associated with high piglet mortality when sows farrow loose housed outdoors in huts but to low piglet mortality when sows farrow loose housed indoors in pens.