The beta-N-methylamino-l-alanine (BMAA) is a non-proteinogenic amino acid discussed to be produced by cyanobacteria forming harmful blooms. Since BMAA is suspected etiological agent in neurodegenerative diseases, there is a need to study and validate whether and in what concentrations can BMAA be present in human tissues. The aim of the present study was to validate analytical and extraction procedures for quantification of non-derivatized BMAA in the urine using liquid chromatography and commercial ELISA Kit. The study was focused on BMAA in different forms - dissolved, protein associated and total. The validated protocol included SPE followed by HILIC MS/MS for analyses of non-derivatized free form of BMAA with a limit of quantification 20 ng/mL. The methods for other BMAA forms (i.e. protein-associated and total) were also assessed but high matrix interferences did not allow their implementation. The method was used for analyses of free BMAA in 23 urine samples from healthy volunteers and psychiatric patients suffering from nonspecific neurological symptoms. Traces of BMAA were suspectedly detected in a single urine sample but they were not unequivocally proved according to all conservative analytical criteria. BMAA was also not confirmed in a repeatedly collected sample from the same person. The evaluated commercial BMAA ELISA Kit (Abraxis) was not suitable for determination of BMAA in extracted urine samples because of systematically highly false positive results. In agreement with recent findings, analyses of BMAA appear to methodologically challenging, and further research on BMAA in human tissues (or its precursors with potency to form BMAA under natural conditions or - eventually - during sample processing) is needed to clarify its potential ethiological role in neurodegenerative diseases.