Application of ORYZA-DSSAT model to estimate upland rice yield as a subsidy to climate risk zoning in the State of Goiás, Brazil Rice is the most important crop for food security, growing in many locations and different climate conditions around the world. In Brazil, rice is grown mainly under flood conditions. Due the increase of rice demand and social and environmental problems caused by the excessive use of water in the flood production system, there has been interest of expanding rice production to other regions of Brazil. In these regions, where rice production system is under rainfed conditions (upland rice), the main factor limiting rice yield is the water stress. To minimize the climatic risks and maximize profits, farmers and government agencies should find out strategies to identify the best sowing dates for upland rice. In such context, the decisions can be done based on crop simulation models, which are very useful tools to predict the variability of yield, defining the best sowing dates. Based on that, the objectives of the present study were: 1) to calibrate and evaluate ORYZA-DSSAT model to estimate the development and yield of upland rice in the State of Goiás, Brazil; 2) to apply the model for determining the best sowing dates, together with the crop water requirement satisfaction index (ISNA) for all phenological phases and production costs, and to compare these dates with those recommended by the Climatic Risk Zoning of Minister of Agriculture, Livestock and Food Supply (MAPA); and 3) to determine the best regions of Goiás State to grown upland rice, thtough yield maps, in order to support the recommendation of rice cultivars and adoption of public policies. To calibrate the model, data from one field experiment carried out in Santo Antônio de Goiás, GO during 2010-2011 season was used. To evaluate the model data from two other independent field experiments were used, with the first carried out in Santo Antônio de Goiás, GO, during 2008-2009 season, and the second in Porangatu, GO, during 2009-2010 season. The upland Brazilian rice cultivar BRS-Primavera (normal season) was the one used in these experiments. The results showed that the model was able to estimate development and yield of upland rice in Goiás State. Differences were found among the best sowing dates determined by this study and those recommended by MAPA. The model ORYZA-DSSAT was efficient for simulating the upland rice potential and attainable upland rice yields in the state of Goiás, in function of temporal and spatial climate variability, making possible to generate maps to subsidize the crop expansion in the state, to allocate the best cultivars to each region and to evaluate the climatic risk of the different sowing dates.