This study focuses on the smart operation of LP gas cylinder distribution. In Japan, many LP gas suppliers utilize Internet of Things technology, primarily for remote monitoring, with approximately half of the customers in Japan being remotely monitored. Daily monitoring of customer inventory via remote metering can significantly reduce both inventory and delivery costs by decreasing the number of visits to customers. In this study, we propose advanced smart operations that leverage seasonal differences in demand. We then formulate the three problems to demonstrate how many cylinders and/or hours a driver saves during peak periods. Because these problems are a mixed-integer nonlinear optimization problem that includes capacity and working time constraints, we propose an iterative procedure to solve the relaxed integer optimization problem and obtain near-optimal solutions. Our results indicate that the maximum load is reduced by 30–40% compared with that of the current distribution system, based on an actual instance of three drivers.