Improving predictions of ponderosa pine (Pinus ponderosa var. scopulorum) regeneration is critical as landscape-scale restoration efforts are being implemented across the western United States. The establishment of new ponderosa pine cohorts following restoration treatments will have long-term impacts on the future development of forest stands. At a regional scale, we used data associated with 364 Forest Inventory and Analysis (FIA) plots within the ponderosa pine forest type on National Forests in Arizona and New Mexico. At a local scale, we sampled 28 stands on basalt and sedimentary parent materials on the Coconino National Forest in northern Arizona. Structural equation models were developed to determine the relative importance of direct and indirect factors affecting ponderosa pine seedling densities at both regional and local scales. The regional model only explained 13% of the variation in pine seedling density. The model indicated that ponderosa pine seedling density was highest where (1) mean annual precipitation was highest, (2) average May minimum temperature was highest, (3) overstory stand basal area was lowest and (4) Muhlenbergia spp. were dominant herbaceous plants. The local model explained 76% of the variation in pine seedling density. Densities were highest where (1) mean annual precipitation was highest, (2) soil clay content was lowest, (3) soil pH was lowest, (4) Muhlenbergia virescens frequency was highest and (5) seed tree frequency was highest. If additional factors such as soil texture, soil pH, screwleaf muhly frequency and seed tree frequency were measured on FIA plots, then perhaps more variation in pine seedling density could be explained at the regional scale.