A total of 2,284 banana unigene sequences were mined, resulting in the identification of 122 SSRs in 110 sequences. Among them, there were 37 motifs the overall average length of SSRs was 20 bp. The trinucleotide repeats appeared to be the most abundant SSRs (47.6%), followed by dinucleotide repeats (33.1%), whereas the tetranucleotide repeats were less abundant. The rich repeats AG and AGG were predominant, accounting for 75.7% and 36% in dri- and tri-nucleotide repeats respectively. A total of 63 primer pairs were designed and tested against genomic DNA of Brazil. Forty-one primers could be validated as usable markers. Nineteen of these primers revealed moderate to high polymorphism information content (PIC) across 49 accessions. The polymorphic bands were ranged from 4 to 12, averagely 7.58. The mean PIC value was 0.7324, ranging from 0.3572 to 0.8744. Forty-nine banana varieties were separated into two major clusters closely corresponding with the genome composition by UPGMA analysis. The results showed that the EST-SSR markers were of great value in evaluation of banana germplasms.