This paper proposes a framework for message transmission scheduling on a simple tandem multi-hop transmission model with lossy unreliable wireless links, where each of N nodes periodically generates a message every T time-slots. Such a model is of practical importance, e.g., in low-cost serially-arranged sensor networks in the wild. Each message can be transmitted to an adjacent in a single time-slot, and should be relayed in a store-wait-and-forward manner from its source node to one of gateways at the edges of the tandem within T time-slots. Our framework consists of: (i) a static global time-slot assignment over all links analytically derived by a central server; and (ii) a local message selection for transmission with a simple XOR network coding-based proactive recovery over assigned time-slots on each link. Simulation results show the probability that all messages are successfully delivered to the server (via gateways) by the proposed framework is comparable to or even better than that by ACK-based reactive recovery schemes.