Female reproductive performance is a central component of ungulate population dynamics, and it can be influenced by individual, social, and environmental factors. Researchers have often assumed direct effects of different predictors on reproduction, yet more complex relationships should be considered when investigating temporal variations in life‐history traits within a broader eco‐evolutionary context. In this study, we explored direct effects of individual, social, and environmental predictors on female reproductive performance and investigated potential causal chains among variables. We analyzed the variation in fecundity, measured as the probability of being pregnant, in 215 adult female deer Cervus elaphus culled on the Italian Alps, with respect to age, body mass, kidney fat, jaw length, lactation status, population size, temperature and precipitation in spring–summer, temperature and snow depth in winter, and the delayed effect of spring–summer temperature. We used random forest and logistic regression models to select variables whose direct effects best explained variation in fecundity. Path analysis was used to test for alternative hypotheses of direct/indirect effects between pre‐selected weather (spring–summer temperature) and individual (age, kidney fat index [KFI]) predictors. The most important direct predictors of fecundity were age, kidney fat, and the interaction between kidney fat and spring–summer temperature. Path analysis supported the hypothesis that higher spring–summer temperature had negative, indirect effects on the probability of being pregnant, mediated by decreasing values of KFI. Our study revealed some complex, cause–effect relationships between weather stochasticity, body condition, and reproduction, possibly suggesting a conditional trade‐off between opportunity for reproduction and survival, and emphasizing how environmental variations and individual characteristics may interact to shape life‐history traits in ungulate populations.