Addressing the issue of delivery efficiency and transportation service quality in rural areas, this thesis presents an analysis of total cost of delivery services in regions with low demand density and low road network density. It focuses on designing a cost-effective and efficient freight delivery system, which is crucial for promoting a vibrant rural economy. A flexible-route service model is developed, aiming to improve farm products and other deliveries by optimizing the service zone size and frequency to minimize the average cost per delivered package. The model is tailored for a potential truck operation scenario in the central Appalachian region, serving as a representative case study, with a general formulation of total cost that can be adapted to similar cases elsewhere. Considering the influence of dead-end roads in rural area, this study presents an adjusted formulation of length estimation for Traveling Salesman Problem (TSP) tours based on the literature review and regression on multiple graphs with road network, and develops a mathematical formulation of total cost, integrating operation and user costs, supported by reasonable assumptions and system constraints. The results from the baseline study suggest that one truck can serve a large service area by exceeding the maximum working hours constraint. This observation is made without considering the potential expansion into a multi-zone system, which might be necessary due to the combined factors of road network complexity and the perishability of farm products. The results from our sensitivity analysis show that a system with a single large truck will have the lowest average cost per package when demand is low. Considering an actual road network, this study also explores the possibility of combining the flexible-route delivery service with self-deliveries and the extension of Vehicle Routing Problem (VRP) with maximum working hour constraint. The study concludes with suggested future research directions in this important domain.