This study, carried out in the Daloa region (Middle-west of Côte d'lvoire), aims at implementing a statistical inventory mapping of coffee and cacao plantations in Côte d'Ivoire by satellite remote sensing. The methodology used resides in classifying a multi-band SPOT image by the "maximum likelihood" method. The first step consisted in an in-depth survey (36 variables, including 19 agronomic ones) carried out on 358 plots in a pilot zone of about 30.000 ha already ground-delimited by Plan Foncier Rural (PFR). Data multivariate analysis revealed that, for both crops, yield estimated by investigators was summing up the key variables describing plantations (shade, condition, maintenance, etc... ). This parameter was thus used to create test-plots samples (coffee and cocoa). The "yield-spatial delimitation of the plots" approach proved to give the best classification results controlled over all studied plots. The test-zone classified image, cartographied with a 1 ha unit is shown. Global mapping precision (Pixels Correctly Classified) is 72 % : 69 % for coffee and 66 % for cacao. In a second phase, the methodology was applied to the entire SPOT scene (360,000 ha) and a validation survey was implemented. Global mapping precision is then 80% and reconnaissance rates for coffee and cocoa are 86 and 94%, respectively. Estimated areas might be more reliable than previous estimations resulting from land or farming investigations used so far. It is worth noting that these results were obtained through reduced test-plots samples (in number-about ten per theme- and in area), hence with low costs and duration in operational phase.