This article describes the development of the Relationship Closeness Inventory (RCI), which draws on the conceptualization of closeness as high interdependence between two people's activities proposed by Kelley et al. (1983). The current "closest" relationship of individuals (N = 241) drawn from the college student population served as the basis for RCI development, with the closest relationship found to encompass several relationship types, including romantic, friend, and family relationships. The development and psychometric properties of the three RCI subscales (Frequency, Diversity, Strength), their scoring, and their combination to form an overall index of closeness are described. The RCI's test-retest reliability is reported and the association between RCI score and the longevity of the relationship is discussed. RCI scores for individuals' closest relationships are contrasted to those of not-close relationships, to a subjective closeness index, and to several measures of relationship affect, including Rubin's (1973) Liking and Loving scales. Finally, the ability of the RCI to predict relationship break up is contrasted to that of the Subjective Closeness Index, an index of the emotional tone of the relationship, and to relationship longevity. It is concluded that the RCI possesses acceptable internal and test-retest reliability, satisfactory discriminant and construct validity, and encouraging predictive validity, thus suggesting that it constitutes a viable means of assessing relationship closeness.