This chapter outlines how landscape simulation models can be used to support forest landscape restoration. In the first type of application, landscape models of disturbance and forest succession were used to estimate historical variability in landscape composition and configuration. An example is given based on a study in the Oregon Coast Range of USA which showed the present day forest patterns are outside the range of historical variability. Problems with this approach lie in deciding the landscape metrics to use and, in this particular case, in assembling reliable data on historical fire regimes. A second common application of landscape simulation models is to project future landscapes under alternative landscape restoration scenarios. These types of simulation experiments with landscape models focus less on making predictions of historical or future landscape conditions but, rather, place more emphasis on exploring general hypotheses about pattern-process relationships. One important insight is that changes in landscape composition and configuration often lag behind shifts in disturbance regimes, and that temporal as well as spatial landscape heterogeneity is important to consider when assessing ecological responses to changing disturbance regimes.