Previous meta-analyses have found small to moderate associations between child-therapist alliance and treatment outcomes. However, these meta-analyses have not taken into account changes in alliance (i.e., alliance shifts), alliance agreement (i.e., congruence or discrepancies between child-therapist ratings), and the role of alliance as a moderator in relation to treatment outcomes (i.e., an interaction effect of alliance and treatment condition on treatment outcomes). A series of multilevel meta-analyses of 99 studies was conducted to investigate several types of alliance-outcome associations in child and adolescent psychotherapy. Associations between child-therapist alliance and child outcomes (r = 0.17), changes in child-therapist alliance and child outcomes (r = 0.19), child-therapist alliance as a moderator of outcomes (r = 0.09), and parent-therapist alliance and child outcomes (r = 0.13) were small. Associations between child-therapist alliance agreement and outcomes (r = 0.21) and between parent-therapist alliance and parent outcomes (r = 0.24) were small to moderate. This meta-analysis provides the most updated and comprehensive overview of the alliance-outcome association in child and adolescent psychotherapy, showing that the alliance continues to show impact on treatment outcomes. Alliance research in youth psychotherapy has increasingly focused on several complex aspects of the alliance-outcome association, such as the role of changes in alliance, alliance discrepancies, client and therapist variability, and the reciprocal association between alliance and prior symptom change in relation to treatment outcomes. Implications for future research and clinical practice are discussed.