The widespread replacement of native species with cosmopolitan, nonnative species is homogenizing the global fauna and flora. While the empirical study of biotic homogenization is substantial and growing, theoretical aspects have yet to be explored. Consequently, the breadth of possible ecological mechanisms that can shape current and future patterns and rates of homogenization remain largely unknown. Here, we develop a conceptual model that describes 14 potential scenarios by which species invasions and/or extinctions can lead to various trajectories of biotic homogenization (increased community similarity) or differentiation (decreased community similarity); we then use a simulation approach to explore the model’s predictions. We found changes in community similarity to vary with the type and number of nonnative and native species, the historical degree of similarity among the communities, and, to a lesser degree, the richness of the recipient communities. Homogenization is greatest when similar species invade communities, causing either no extinction or differential extinction of native species. The model predictions are consistent with current empirical data for fish, bird, and plant communities and therefore may represent the dominant mechanisms of contemporary homogenization. We present a unifying model illustrating how the balance between invading and extinct species dictates the outcome of biotic homogenization. We conclude by discussing a number of critical but largely unrecognized issues that bear on the empirical study of biotic homogenization, including the importance of spatial scale, temporal scale, and data resolution. We argue that the study of biotic homogenization needs to be placed in a more mechanistic and predictive framework in order for studies to provide adequate guidance in conservation efforts to maintain regional distinctness of the global biota.