In order to study the molecular mechanisms underlying the response to drought stress in coffee plants, meristems of plagiotropic branches from Iapar59 (I59, drought tolerant) and Rubi (R, drought susceptible) cultivars of Coffea arabica grown in field-grown conditions and submitted (NI: non-irrigated) or not (I: irrigated) to water stress were collected at the end of the dry season and used to generate cDNAs that were sequenced using the GS-FLX Titanium strategy. The crude analysis of sequencing data revealed 282.213, 135.304, 345.751 and 230.064 reads obtained from I59-I, I59-NI, R-I and R-NI samples, respectively, totalizing more than 255Mb. For both experiments, most of the reads (>65%) had high quality and were assembled to generate around 24.000 contigs with an average length above 500pb. These data, which are the first one coffee meristems subjected to water stress, will be compared with all available transcriptome Coffea data, including the EST sequences from both Coffea arabica and C canephora. The results of these comparisons will be presented as well as the preliminary data of an electronic northern performed in order to identify differentially expressed genes between the two cultivars in both conditions (NI or I).