Teacher consultation is commonly used to ensure that classroom-based interventions are implemented with fidelity to achieve targeted outcomes, yet the consultation process is not well-understood. Consultant-teacher relationship quality is one feature of consultation that may promote intervention outcomes--both directly and indirectly via teachers' implementation. The current study used mediation models to examine the direct links between consultant-teacher relationship quality, assessed from the perspective of the consultant and teacher, and observed dyadic teacher-child interactions, as well as the indirect effect through teachers' implementation fidelity and dosage. Implementation data come from the Banking Time intervention (N =168 children, 57 teachers), a dyadic teacher-child intervention that targets the quality of interactions between teachers and preschoolers perceived to display disruptive behavior. Consultants (N = 4) worked with teachers to support their implementation of specific Banking Time practices. Findings supported a direct link between consultant-reported relationship quality and teachers' observed interactions with children, however, no evidence for an indirect effect was found. Consultant-reported relationship quality predicted implementation dosage but not fidelity. Across findings, consultant-reported relationship quality emerged as a stronger predictor of outcomes than teacher-reported relationship quality. Results have implications for school-based interventions that employ teacher consultation to support teachers. [This paper was published online in "School Mental Health" 2021.]