Aim. Lactate minimum test (LMT) has become popular in running evaluation. This study analysed the influence of different stage lengths and determination methods on lactate minimum swimming speed (LMS) and its validity for maximal lactate steady-state speed (MLSS-S) assessment. Methods. Twelve male swimmers (19.7+/-1.6 years, 70.3+/-8.7 kg, 181.4+/-7.9 cm) randomly underwent 4-5 evaluations in a 2-week period. LMS was accessed by simple visualisation (SV) and spline function (SF) methods during 200 and 300 in stages LMT (LMT200 and LMT300, respectively), and MLSS-S was determined during constant speed 2000 m efforts. Results. Respectively, SV and SF provided LMS during LMT200 (1.31+/-0.12 m.s(-1) and 1.32+/-0.10 m.sec(-1)) and LMT300 (1.28+/-0.11 m.sec(-1) and 1.28+/-0.10 m.sec(-1)) which were not significantly different (p>0.05) from each other. However, LMS accessed during LMT200 were significantly greater (p<0.05) than MLSS-S (1.25+/-0.06 m.sec(-1)). In addition, significant relationships (r=0.79 to 0.98; p<0.05) were found between all studied speeds and lactate minimum values during LMT300 were not significantly different (p>0.05) from those found during LMT200 (5.4+/-2.2 and 5.5+/-2.2 mM vs 6.8+/-2.5 and 7.0+/-2.6 mM, respectively for SV and SF). Conclusions. Our results suggest that LMS is not affected by different stage lengths and determination methods. However, LMT300 results seems to provide a more accurate MLSS-S prediction, being adequate for swimmers evaluation.433312318