The advancement of next generation sequencing technologies (NGS) has revolutionized our ability to generate large quantities of data at a genomic scale. Despite great challenges, these new sequencing technologies have empowered scientists to explore various relevant biological questions on non-model organisms, even in the absence of a complete sequenced reference genome. Here, we analyzed whole flower transcriptome libraries from exemplar species across the monocot order Zingiberales, using a comparative approach in order to gain insight into the evolution of the molecular mechanisms underlying flower development in the group. We identified 4,153 coding genes shared by all floral transcriptomes analyzed, and 1,748 genes that are only retrieved in the Zingiberales. We also identified 666 genes that are unique to the ginger lineage, and 2,001 that are only found in the banana group, while in the outgroup species Dichorisandra thyrsiflora J.C. Mikan (Commelinaceae) we retrieved 2,686 unique genes. It is possible that some of these genes underlie lineage-specific molecular mechanisms of floral diversification. We further discuss the nature of these lineage-specific datasets, emphasizing conserved and unique molecular processes with special emphasis in the Zingiberales. We also briefly discuss the strengths and shortcomings of de novo assembly for the study of developmental processes across divergent taxa from a particular order. Although this comparison is based exclusively on coding genes, with particular emphasis in transcription factors, we believe that the careful study of other regulatory mechanisms, such as non-coding RNAs, might reveal new levels of complexity, which were not explored in this work.