Phthalic acid diesters (PAE) are omnipresent in the human environment and food is a major contributor to the overall human exposure towards these chemicals. Due to developmental effects, PAE infants' exposure via human milk has been subjected to a number of analytical studies. These previous studies, however, revealed that normal laboratory blank values are in the range of or even higher than human milk levels due to the presence of PAE in laboratory environments. In order to provide more reliable data on PAE exposure via human milk, the aim of this study was to develop and validate a robust and sensitive analytical method. This should be capable of removing matrix components efficiently and guarantee limits of quantification in the low ppb range. The method development took into account liquid-liquid extraction and selective pressurised fluid extraction (sPFE) as well as chromatography-based clean-up steps. The final method consisted of a liquid-liquid extraction followed by an automated chromatographic clean-up by an sPFE device. After volume reduction the cleaned extracts were analysed by quadrupole GC/MS. Quantification was based on internal standards. An extensive quality assurance and method test programme demonstrated conservatively determined limits of detection and quantification from 0.3 to 10 ng g⁻¹ in human milk, with recoveries of internal standards from 50% to 101%. Thus, the method allowed the quality-assured detection of di-isobutyl phthalate (DiBP), di-n-butyl phthalate (DBP), di-2-ethylhexyl phthalate (DEHP), di-allyl phthalate (DAP), benzylbutyl phthalate (BBP) and di-cyclohexyl phthalate (DcHP) in 30 human milk samples provided by 30 volunteers from southern Germany. DiBP, DBP and DEHP were the most commonly detected PAE, with median levels of 1.0, 0.6 and 2.3 ng g⁻¹, respectively.