Purpose We compared the pre- and post-irradiation viability and cytotoxicity of human peripheral natural killer cell (NK) populations obtained using different isolation methods. Material and methods Three methods were used to enrich total NK cells from buffy coats: (I) a Ficoll-Paque gradient, plastic adherence and a nylon wool column; (II) a discontinuous Percoll gradient; or (III) the Dynal NK cell isolation kit. Subsequently, CD16(+) and CD56(+) NK cell subsets were collected using (IV) flow cytometry or (V) magnetic-activated cell sorting (MACS) NK cell isolation kits. The yield, viability, purity and cytotoxicity of the NK cell populations were measured using trypan blue exclusion, flow cytometry using propidium iodide and (51)Cr release assays after enrichments as well as viability and cytotoxicity after a single radiation dose. Results The purity of the preparations, as measured by the CD16(+) and CD56(+) cell content, was equally good between methods I-III (p = 0.323), but the content of CD16(+) and CD56(+) cells using these methods was significantly lower than that using methods IV and V (p = 0.005). The viability of the cell population enriched via flow cytometry (85.5%) was significantly lower than that enriched via other methods (99.4-98.0%, p = 0.003). The cytotoxicity of NK cells enriched using methods I-III was significantly higher than that of NK cells enriched using methods IV and V (p = 0.000). In vitro the NK cells did not recover cytotoxic activity following irradiation. In addition, we detected considerable inter-individual variation in yield, cytotoxicity and radiation sensitivity between the NK cells collected from different human donors. Conclusions The selection of the appropriate NK cell enrichment method is very important for NK cell irradiation studies. According to our results, the Dynal and MACS NK isolation kits best retained the killing capacity and the viability of irradiated NK cells.