The sequencing of the ESTS of three coffee species transcriptome, Coffea arabica, Coffea canephora and Coffea racemosa, opened many possibilities for studying different characteristics between these species. Fenotipical characteristics of agronomic interest can be molecularly studied through the transcript analysis of these three species and also through the mapping of a set of target genes for genetic improvement programs. In this work, a clustering for similarity of ESTs of the three species was carried, aiming a study of single nucleotide polymorphisms (SNPs). Using parameters of similarity (>95%) and overlapping (>100 bp) 15,885 contigs have been constructed, which contain a frequency of 0,47 SNPs for 100 bp. The results were stored for consults and future studies of variability in individuals of each species and between these three species.