MicroRNAs (miRNAs) play a crucial role in the post-transcriptional regulation of eukaryotic gene expression, in plants and animals. Many aspects of the biogenesis and evolution of miRNAs in animals and plants differ. For example, unlike miRNAs of animals, which are mainly found in introns or exons from coding genes, most plant miRNAs are encoded by discrete genes. Moreover, miRNAs are released from their precursors using distinct pathways in the two kingdoms. Also, miRNA precursors are more heterogeneous in plants than in animals, varying greatly in size and structure. These differences have justified dedicated approaches for miRNA gene finding. However although several prediction tools are available for metazoan genomes, the number of tools dedicated to plants is relatively limited. Considering this gap, we have developed miRkwood, a user-friendly web server specifically designed for plant miRNAs. miRkwood is able to face the diversity of plant pre-miRNAs and allows the prediction of precursors of both conserved and non-conserved miRNAs. Moreover, it offers an intuitive and comprehensive user interface to navigate in the data, as well as many export options (GFF, CSV, FASTA, ODT) to allow the user to conduct further analyses on a local computer. miRkwood can deal with both short genomic sequences (up to 100 000 nt) and full small RNA sequencing reads (provided that a BAM file is available). It is accessible at http://bioinfo.lifl.fr/mirkwood