Contemporary forest management often consists of multiple objectives, including restoration of human-impacted forested landscapes toward their range of natural variability (RNV) and sustainable levels of timber production. Balancing multiple management objectives is often challenging due to intrinsic conflicts between these objectives and a lack of reference conditions for evaluating the effectiveness of forest restoration efforts. We used a spatially explicit forest landscape model to assess how well a classification-based forest management (CFM) system could achieve multiple objectives in a Korean pine broadleaf mixed forest ecosystem at Changbai Mountain in Northeast China. The CFM system divided the forest landscape into three management areas (Commercial Forest, Special Ecological Welfare Forest, and General Ecological Welfare Forest), each with its own management objectives and prescriptions, but with an overall goal of increasing the ecological and economic sustainability of the entire landscape. The zoning approach adopted in the Chinese CFM system is very similar to the TRIAD approach that is being advocated for managing public forests in Canada. In this study, a natural disturbance scenario and seven harvest scenarios (one identical to the current harvest regime and six alternative scenarios) were simulated to examine how tree species composition, age structure, and timber production at the landscape level can be affected by different strategies under the CFM system. The results indicated that the current forest management regime would not only fail to reach the designated timber production level but also move the forest landscape far away from its RNV. In order to return the currently altered forest landscape to approach its RNV while providing a stable level of timber production over time, harvest intensities should be reduced to a level that is equivalent to the amount of timber removals that would occur under the natural disturbances; and the establishment of forest plantations is also required.