Manufacturing firms are often faced with frequent price fluctuations of both raw materials and the final products. Therefore, it has become increasingly difficult for the firms to make proper procurement and production decisions. This paper studies the inventory and production decisions for a manufacturer who is faced with fluctuating material costs. A multi-period inventory and production decision model is developed based on the assumptions that the fluctuation of material costs follows the Markov process and the market demand for the final product is a price-dependent random variable. The optimal dynamic inventory and production policies are then analyzed. One unique feature of the model is that the manufacturer can sell raw materials. Meanwhile, the manufacturer has to make inventory decisions of the final product. It is shown that the optimal production and inventory decisions are of base-stock type. Furthermore, several structural properties are developed to further characterize the detailed policy structure. Extensive numerical studies verify the optimal policies and demonstrate how the model parameters affect the optimal decisions.