We explain ‘Naiji ordering system’ between an automotive manufacturer and suppliers. ‘Naiji’means ‘forecast order’ which is transmitted from a manufacturer in advance before suppliers' manufacturing. We
focus on how to make supplier's weekly production plan with a resilience constraint. Resilience is powers of
recovery against risk or business continuity. And we research on how to track the risk and on how many volume for an inventory manager to fulfill manufacturer's orders. We propose a risk measure for resilience. It is based on an expected stock-out for planning horizon and we prove the characteristics of the measure. We propose a supplier multi-period inventory control model with a resilience constraint. It is formulated as a dynamic programming problem. We show the availability of the model by numerical experiment.