Fire and timber harvest are two major forest disturbances in boreal forests. Predicting the dynamics of boreal forest biomass requires accounting for both of those effects. Related stochasticity and other uncertainties can produce great variation in predicted responses of forests to fire and timber harvest. In this study, we investigated the effects of fire and timber harvest on landscape-level predictions of the tree component of stand biomass in a boreal forest landscape in Northeast China. We used a forest landscape model (LANDIS PRO) to predict the tree biomass over three time intervals (0-50, 50-150, and 150-300 years). We then compared the simulated results of fire and timber harvest and their interactions with observed biomass and its spatial distribution over short-, mid-, and long-term intervals. For additional prediction comparisons, we observed uncut, unburned stands (i.e., the succession-only scenario). Compared to the succession-only scenario, we found that predicted biomass was reduced by 3.8 +/- A 2.1, 9.1 +/- A 3.6, and 11.2 +/- A 5.1 tons/ha in fire-only, harvest-only, and combined fire and harvest scenarios, respectively. Our results indicated that the effect of harvest on biomass exceeded that of fire, and that the interaction of fire and harvest was more effective in reducing biomass than the effects of fire or harvest separately. Biomass predictions that did not consider effects of fire and timber harvest tended to inflate biomass estimates. The spatial distribution of tree biomass moreover changed with simulation period. These results have important implications in designing prescriptions for improving forest sustainability.