Transcriptomic sequence resources represent invaluable assets for research, in particular for non-model species without a sequenced genome. To date, the Next Generation Sequencing technologies 454/Roche and Illumina have been used to generate transcriptome sequence databases by mRNA-Seq for more than fifty different plant species. While some of the databases were successfully used for downstream applications, such as proteomics, the assembly parameters indicate that the assemblies do not yet accurately reflect the actual plant transcriptomes. Two different assembly strategies have been used, overlap consensus based assemblers for long reads and Eulerian path/de Bruijn graph assembler for short reads. In this review, we discuss the challenges and solutions to the transcriptome assembly problem. A list of quality control parameters and the necessary scripts to produce them are provided.