Objectives: Methamphetamine (METH) use has been associated with impairments in a variety of cognitive functions. In this study, it was aimed to assess the relation between cognitive measures and frequency of METH use. Methods: Ninety-eight participants with methamphetamine use disorder were assessed with Wisconsin Card Sorting Test (WCST), Victoria-Stroop Word Color Test (SWCT), Trail Making Test (TMT) A and B. Participants were followed up during 6-month period with multiple urine toxicology tests (at baseline, month 1, month 2, month 3 and month 6). Results: Participants who had their prescriptions regularly were more likely to have higher rates of METH-negative urine results ( P = .003). Participants who had more correct numbers, completed more categories and had more conceptual level responses on WCST were more likely to use lower frequency of METH (OR = 0.006, P < .01; OR = 0.171, P < .001; OR = 0.024, P < .001; respectively). The higher error numbers and perseverative error rates on WCST were associated with more frequent METH use (OR = 0.023, P < .001; OR = 0.076, P < .001). Interference factor on SWCT was related with lower frequency of METH use whereas color naming factor on SWCT was related with higher rates of urine results (OR = 0.012, P < .001; OR = 3.628, P < .001; respectively). Higher TMT B-A score was related with more frequent METH use, although the significance disappeared after adjustment (OR = 0.002, P < .001). Having psychotic symptoms predicted less frequent use, but after adjustment for other significant variables, there was no significance. Conclusion: Lower frequency of METH use in the follow-up could be predicted by neurocognitive assessments. Deficits in executive functions, attention, set-shifting and mental flexibility seems to be the most effected domains and this effect may be independent from the severity of psychotic symptoms.