By overcoming the distance limit in quantum key distribution (QKD) technology, key relaying over trusted nodes is regarded as an essential function for practical quantum-secured networking. The key relay function consumes a series of costly quantum key resources to establish an end-to-end key in QKD networks (QKDNs); thus, an efficient key relay algorithm is required. To investigate this problem, we developed an integer linear programming (ILP) formulation that maximizes a balanced summation between max-min fairness and efficiency in the key relay. Inspired by the ILP model, this study developed a key-count differential-based proactive key relay (DPKR) heuristic algorithm to provide a scalable solution for key relays in QKDNs. The proposed DPKR algorithm iteratively selects a set of nodes by considering the key-count differential between nodes and establishes an end-to-end key by relaying it over the selected nodes. Owing to the proactive key relay model, an effective key relay route in the key management layer in the QKDN can be calculated solely by the key-count status in the key management layer. The proposed DPKR algorithm reduces manifold order-of-magnitude in the algorithm computation time at the cost of an acceptable loss in the average and minimum numbers of keys from those of the ILP optimization model. The computation time evaluation clearly manifests a scalability of the DPKR algorithm in quantum-secured networking.