When wind power constitutes a larger share of the electricity production mix, credible and reliable modelling of its operation in long-term investment models becomes increasingly important. In this paper the intermittent characteristics of wind power are modelled as a stochastic parameter in a long-term TIMES model of the Danish heat and electricity sector. To our knowledge, this is not a common approach in long-term investment models, and has not been done previously in TIMES, where the short-term uncertainty of wind power is normally taken into account by a deterministic constraint that ensures excess back-up capacity. In our model, the stochasticity gives lower total energy system costs, significant lower investments in wind power, less expected electricity export and higher expected biomass consumption compared to using the traditional deterministic approach. Also, the deterministic investment strategy can be insufficient in periods with poor wind conditions. Based on our findings, we recommend using a stochastic representation of intermittent renewables in long-term investment models to provide more solid results for decision makers.