Reconfigurable battery systems (RBSs) provide a promising alternative to traditional battery systems due to their flexible and dynamically changeable topological structures that can be adapted to different battery charging and discharging strategies. A critical system parameter known as the maximum allowable current (MAC) is pivotal to RBS operation. This parameter is instrumental in maintaining the current of each individual battery within a safe range and serves as a guiding indicator for the system’s reconfiguration, ensuring its safety and reliability. This paper proposes a method for calculating the MAC of an arbitrary RBS using a greedy algorithm in conjunction with a directed graph model of the RBS. Using the shortest path of the battery, the greedy algorithm transforms the exhaustion of the switch states in the brute-force algorithm or variable search without utilizing structures in the heuristic algorithms in the combination of the shortest paths. The directed graph model, based on an equivalent circuit, provides a specific method for calculating the MAC of a given structure. The proposed method is validated using 2 previously published RBS structures and an additional one with a more complex structure. The results are the same as those from the brute-force algorithm, but the proposed method substantially improves the computational efficiency, being theoretically N s 2 N s − N b log 10 N b times faster than the brute-force algorithm for an RBS with N b batteries and N s switches. Another advantage of the proposed method is its ability to calculate the MAC of RBSs with arbitrary structures and variable batteries, even in scenarios with random isolated batteries.