The aim of this study was to develop a multiple-level-of-analysis model of preschool attachment security and to determine the processes (direct and indirect) whereby factors from different domains (e.g., stress and parenting) are related to attachment during this period. This study examined the direct and indirect effects of stress, family conflict, caregiver depression symptoms, and parenting on attachment security in a large (N = 796) and diverse sample of 4-year-olds. This study used the 3-Boxes Task to assess aspects of parenting critical to sensitivity in the preschool period, labeling this construct sensitivity/scaffolding. Parent-report questionnaires were used to assess stress, conflict, caregiver depressive symptoms, parent support/engagement, and parent hostility/coercion. Direct observation (3-Boxes Task) was used to assess sensitivity/scaffolding and attachment (Attachment Q-Sort) based on a 2(1/2)-3 hour home visit. Results of structural equation modeling indicated a good overall fit for the model. Among the parenting variables, sensitivity/scaffolding had the strongest effect on attachment. Depressive symptoms had both direct and indirect effects (mediated by parenting). The effects of stress and family conflict were mediated by caregiver depression symptoms and parenting. These data show that a developmentally appropriate measure of sensitivity plays a significant role in attachment security in preschoolers. Thus, strategies designed to enhance sensitivity/scaffolding may increase child resilience by enhancing attachment security.