Femtocells have been assuming an increasingly important role to improve the coverage and capacity of cellular networks over the next few years. Different from macrocellular networks, femtocells, which are run by an operator, can provide three access modes, i.e., open, closed, and hybrid access. As a compromise of both open and closed access mode, hybrid mode is considered the most efficient for both operators and customers. In hybrid access mode, a femtocell normally prioritizes different types of users, and users always prefer to choose the femtocell with the best signal quality for better service and saving power, giving rise to a unique mutual preference. This unique mutual preference relationship between femtocells and users makes the admission control much more challenging, and no current approach considers this mutual preference when hybrid femtocells are accessed by many users. In this paper, we model a mutual selection admission (MSA) problem for hybrid femtocells and users, aiming to find a stable matching, where both the preference of femtocells and that of users are satisfied under a backhaul capacity constraint. We show that the problem of seeking an optimal stable matching with maximum cardinality is NP-hard, and we propose an efficient heuristic algorithm for solving the MSA problem by seeking a matching with different levels of stability. With extensive simulations, we show that the MSA algorithm can always find a stable matching and have a great rate gain compared with hybrid access for a wide range of users and femtocells. © 1967-2012 IEEE.