Recently, the Cell-Free Massive Multiple-Input Multiple-Out-put (MIMO) architecture has emerged as a promising solution for future wireless communication systems, where a substantial number of distributed wireless Access Points (APs) concurrently serve a significantly smaller count of User Equipment (UE). In this paper, we study the AP selection problem in user-centric cell-free massive MIMO, where each user is served by a restricted number of APs. To address this problem, we propose a Branch-And-Bound (BAB)-based AP selection algorithm to achieve maximum channel capacity, which is designed to efficiently obtain the optimal subset of APs for each user. Our simulation results show that the proposed algorithm outperforms other baseline methods in terms of channel capacity at the expense of some complexity. Meanwhile, our complexity is much lower than the exhaustive search, which also yields optimal results.