Over the last few decades, inventory management of produced, recovered, and returned items in the closed loop supply chain (CLSC) systems has been significantly taken into consideration. Most of available works in the literature assumed that manufactured and remanufactured items have the same quality. However in practices, some customers perceive remanufactured items as lower quality products rather than newly produced items. Therefore, this paper assumes demands for manufactured items and remanufactured ones are not same. This assumption causes lost sales in which there are shortage periods for both produced and reproduced items. In the real-life production systems, the generation of nonconforming items is inevitable owing to deteriorating process or other uncontrollable factors. In order to take this issue into account, the production and reproduction processes are assumed to be imperfect as it can produce some defectives which would be reworked in the same cycles. Moreover, to make the study realistic, we assume the return rate of used items as a variable function which depends on the minimum acceptable quality level of returns. Furthermore, in this research the quality of returned items is considered as a random variable and based on this assumption, buyback cost, remanufacturing cost and salvage value depend on quality level of returned items. Basically, three different stochastic models, with different probability distribution functions, are mathematically developed for multiple manufacturing and remanufacturing cycles. A solution procedure is introduced and then there numerical studies are provided to show the applicability of the developed models and also to perform a sensitivity analysis. Finally, conclusion and future researches are presented.