In Uzbekistan cotton production substantially contributes to GDP. The cotton is produced based on the cotton procurement policy, according to which farmers have to allocate half of their land for cotton and produce certain amount of cotton. However, cotton yields are uncertain. Lower than expected cotton production by farms can be considered as inefficient farm. Because farmers lease land from the state, failure to deliver the cotton output lead to adjustment in its scale of operations – a process called farm optimization. Reduction in farm size of one farmer leads that the area of more efficient farmer that accomplished the cotton production increases. The study aims to develop policies that can improve cotton production and farm incomes. For this we developed a dynamic recursive model that considers variability and farm adjustments. We showed that variability influences farm sizes and due to farm size changes the income inequality may widen among farms.