Irrigation water management is crucial for agricultural production and livelihood security in Morocco as in many other parts of the world. For the implementation of an effective water management knowledge about farmers’ irrigation water demand is crucial to assess demand reactions of a water pricing policy, to establish a cost-benefit analysis of water supply investments or to determine the optimal water allocation between different users. Previously used econometric methods providing this information often have prohibitive data requirements. In this paper, the Contingent Valuation Method (CVM) is adjusted to derive a demand function for irrigation water along farmers’ willingness to pay for one additional unit of surface water or groundwater. An application in the Middle Drâa Valley in Morocco shows that the method provides reasonable results in an environment with limited data availability. For analyzing the censored survey data, the Least Absolute Deviation estimator was found to be a suitable alternative to the Tobit model when errors are heteroscedastic and non-normally distributed. The adjusted CVM to derive demand functions is especially attractive for water scarce countries where water management doubtlessly plays a decisive role.