Understanding how horses interact with landscapes is key to designing environments that support welfare and biodiversity. Yet, little is known about how domestic horses use specific elements within a landscape. This study examined the behavioral interactions of seven Swedish warmblood mares (1–3 years old) with naturally occurring landscape features in the Kumlan Nature Reserve (Sweden) during July and August 2021. HoofStep® sensors continuously recorded equine behaviors, categorized as highly active, active, resting, or eating. Manly Selection Ratios (MSRs) were used to assess landscape feature selection (LFS) relative to availability. A generalized linear model (Gamma distribution, log link) tested the effects of horse, behavior, landscape feature, time of day, temperature, rainfall, and month on LFS. Significant main effects included horse, landscape feature, month, rainfall, and temperature (p < .001). Two-way interactions showed that behavior was linked to LFS and that selection was influenced by weather. For instance, tree rows and hedges were preferred during rainfall (Exp(B) = 1.17, p = 0.01), but avoided as temperatures rose (Exp(B) = -0.51, p < 0.001). Three-way interactions highlighted individual preferences, i.e., Horse A preferred resting on a sandbank (Exp(B) = 42.48, p < 0.05), and Horse B in a blackberry patch (Exp(B) = 25.22, p < 0.05). Horse C was active on a sandbank with vegetation (Exp(B) = 22.57, p = 0.01), while Horse D preferred the pool (Exp(B) = 90.44, p < 0.001). Findings suggest that both landscape and weather shape equine behavior, with notable individual variation. Landscape design should incorporate diverse features to meet the behavioral needs of individual horses and to support welfare and biodiversity goals.