Are there any economic policies able to break the vicious circle of the poverty trap in rural areas of poor countries such as Burkina Faso? To help answer this question, a detailed analysis of production systems has been used to elaborate dynamic household models. A mathematical programming model represents the economic decisions of farmers for the three farm types found in the Plateau Central region. Farms are linked together by factor markets. The model calculates a wide range of variables: incomes of agricultural households and their sources, land allocation between the different crops, cropping techniques used, agricultural production, consumption expenditure and labor allocation among the various activities including off-farm activities. These estimates can be obtained by farmer groups and aggregated for the Plateau Central. The model is used to simulate the effects of five economic policies and their combinations: (i) increasing the availability of irrigation water, (ii) lowering marketing costs, (iii) access to animal traction, (iv) access to credit, (v) reducing producer price variability. The analysis reveals different impacts in function of the different farm groups. Public good policies benefit all groups whereas policies aiming at improving capital market access or at diminishing risks on product sales primarily benefit the richest group. Policies focusing on access to equipment are of greater benefit to the poorest. Given the depth of initial poverty, it is necessary to combine all the measures to allow the incomes of the poorest group to cross the poverty line.