AbstractFrom the background of sustainable development of forestry, biodiversity in forest should be maintained and protected when managing an ecosystem. However, as a result of the complex meaning of biodiversity, ordinary indices used in ecology and forestry are not commonly used by forest managers because of their high requirements of data and precision in order to convey so much information on biodiversity in a single index. On the basis of recent research on virgin forests around the world, Franklin advocates that diversity of forest stand structure could be used as an indirect indicator of forest biodiversity, and this has been accepted by many forest experts. As a result, some new ideas on how to manage forest ecosystems through structure control or restructuring have been put forward. In order to manage forest ecosystem through stand structure control, the significance of forest structure must be clarified and some new indices must be created for use in forest survey and in proper forest measurements. In this study, a set of structural diversity indices for broadleaved Korean pine forest were put forward and tested in different types of stands, including virgin forest, forest after selection cutting, and forest regenerated after clear cutting. Based on the measurement of stratified coverage in addition to plotless sampling, the vertical structure index (VSI ) and horizontal heterogeneity index ( HHI ) were generated, along with descriptions of other structural elements such as downed logs, standing poles and canopy gaps. The basic meaning of VSI is the volume of space occupied by branches and leaves, which was calculated by coverage in each layer and its weight. The basic meaning of HHI is the difference of coverage in all layers between locations in the stands, which was calculated as a community dissimilarity index. The bigger the two indices, the more habitat types for living organisms exist in the forest. At the same time, other ancillary indices, such as species composition, amount of coarse woody debris and gaps were surveyed through plotless sampling methods and transects with plots. Comparison between different stands of virgin forest indicates that these indices were precise enough to describe the spatial structure of the stands and were in agreement with analyses resulting from common forestry survey methods. Used in secondary forest stands, they could also provide more information than common single biodiversity indices. Before using data from plotless sampling, the similarity coefficients between plot sampling and plotless sampling were calculated. The precision of the two methods proved to be identical. Used in stands with different disturbance histories, these indices illustrated remarkable differences in the structures of the stands, indicating that the indices are suitable for forest survey and for guiding choice of management measures in forest ecosystem management. Disadvantages and methods for improvement of the indices are discussed and research priorities are proposed.