Technical Abstract: A new method based on the analysis of mitochondrial intergenic regions characterized by intraspecific variation in DNA sequences was developed and applied to the study of the plant pathogen Phytophthora nicotianae. Two regions flanked by genes trny and rns and trnw and cox2 were identified by comparing the whole mitochondrial genomes of P. infestans, P. ramorum and P. sojae and were amplified using primers designed from the flanking conserved genes. These regions were sequenced from 51 isolates of P. nicotianae of both A1 and A2 mating type recovered from different hosts and geographic regions. Amplicon length varied from 429 to 443 bp (trny/rns) and 322 to 373 bp (trnw/cox2) with intraspecific variation due to single nucleotide polymorphisms and indels. Seventeen, 7 and 20 different haplotypes were detected by individually analyzing regions trny-rns, trnw-cox2 and the combined data set of sequences from both regions, respectively. Phylogenetic analysis inferred with 3 different methods enabled the grouping of isolates in 5 clades, each containing different mitochondrial haplotypes and revealed diversity in the mitochondrial genome of P. nicotianae. The majority of isolates from citrus grouped in a single clade indicating either movement of isolates on planting stock or an association of particular isolates with this host. Phylogenetic groups were not correlated with the radial growth rate of the isolates or the rapidity of apple flesh colonization. The method developed in the present study represents an innovative molecular tool for the characterization of natural populations of P. nicotianae and should be easily expanded to other species of Phytophthora as well as other plant pathogens. It can be used to track specific haplotypes and, thanks to its high genetic resolution, it could be standardized and applied in a DNA barcoding like strategy for the precise identification of sub-specific taxa. Compared to alternative molecular methods, a major advantage is that results are unbiased (a list of nucleotides) and highly reproducible, thus enabling the comparison of data from different laboratories and time periods. Furthermore, the method could be further implemented by the identification of additional variable mitochondrial and/or nuclear genomic regions.