The aim of the study was to address the question of whether shelterwood cutting should be discouraged as a forest regeneration method in Norway spruce(Picea abies(L.) Karst.) forests on highly productive peatlands due to a higher risk of windthrow in such forests. The total extent of windthrow was observed during six years after shelterwood cutting in nine spruce forests on fertile peatlands in north, central and south Sweden. Shelterwoods at densities of 140 and 200 stems ha‐1were studied. At the end of the six year study period, 43% of all trees in the sparser shelterwoods and 38% of the trees in the denser shelterwoods had blown down. With respect to results from recent studies of the development of naturally regenerated and planted seedlings under the shelter trees, the extent of windthrow was regarded acceptable. Analysis of wind data from national weather stations close to the experiments showed that the acceptable extent of windthrow could not be explained by low frequencies of high winds during the shelterwood period. On the contrary, according to the wind observations made at the weather stations selected in central and south Sweden, the annual number of days with high winds (>21 ms‐1) during the six years after the shelter‐wood cutting was larger than the average high wind frequency for the last 40 years. Observations of maximum wind speeds show that the mean annual gale force during the shelterwood period was higher than the long term average. Thus, the results from this study do not justify avoidance of shelterwood cutting in spruce peatland forests because of the risk of windthrow. Each of the shelterwoods in the experiment was located close to a large clearcut, and the shelterwood cuttings were carried out as heavy thinnings. For further reduction of windthrow, denser shelterwoods and no clearcutting in adjacent stands are suggested when using shelterwood systems in practical forestry.