A viscoelastic fictitious crack model has been applied in this paper to the analysis of large‐scale in situ sea ice tests. By matching the experimental response, a stress‐separation curve has been derived for the in situ response of sea ice. The importance of including time dependent deformations in sea ice fracture models is illustrated. The limitations of the fictitious crack model in reproducing the ‘stick‐slip’ crack growth typically observed in sea ice are revealed. Copyright © 1999 John Wiley & Sons, Ltd.