Abstract: Managing forest ecosystems for sustainable, multiple use requires forest resource managers to understand how species composition and distribution vary across environmental gradients and respond to landscape scale disturbance. A number of statistical modeling tools are available to construct predictive models and maps from response data, a set of predictor variables, and a predefined statistical distribution. Non-Parametric Multiplicative Regression (NPMR) is a probability modeling system that finds the best multiplicative set of predictor variables. The best set maximizes the Bayes Factor value which is a ratio based on modeled estimates and a species' average frequency of occurrence. This study demonstrates predictive vegetation modeling and mapping using NPMR and species presence/absence data collected from 610 plots located across an industrial managed forest landscape in Northeast Oregon. Plots were stratified with a random sampling design. Four modeling approaches were taken to compare the predictive power of spatial coordinates in combination with a set of topographically-derived and stand structural predictor variables. Spatial coordinates were often the most powerful predictors and the modeling approach with physiographic and stand structural variables together was frequently the most improved relative to the average frequency of occurrence. Comparisons between Logistic Regression (LR) and NPMR models were conducted for the species Clintonia uniflora (CLUN) and Pinus ponderosa (PIPO). NPMR performed better for CLUN when the best predictor variables selected by NPMR were used to construct a LR model. For PIPO, the performance of NPMR was comparable to LR when the set of predictor variables used to build the LR model was based on whether the response in probability to each variable was monotonic. Species-level GIS probability maps were produced with the application of the physiographic models and a corresponding set of GIS raster files. GIS overlays of indicator species maps were used to construct plant association group (PAG) maps. Intersections of PAG layers resulted in quantitative mapping of intergrade between types. PAG layers were often significant predictor variables in probability models for 70 understory and five conifer species produced with Logistic Regression (LR) using a forward step-wise process. Potential applications of both NPMR and LR models with the Forest Vegetation Simulator are discussed.