An interaction between smoking and antipsychotic medications could potentially affect treatment efficacy and promote metabolic side effects. We investigated the contribution of smoking status towards Positive and Negative Syndrome Scale (PANSS) scores and metabolic syndrome- related parameters (plasma lipid and glucose concentrations, and body mass index) among two groups of unmedicated patients with psychosis from the Croatian population: antipsychotic-naïve first-episode patients and nonadherent chronic patients. Previous data are inconsistent regarding the effects of smoking on clinical psychopathology among antipsychotic-naïve or minimally medicated patients with first-episode psychosis, and no studies have examined the potential influence of smoking on clinical psychopathology and metabolic parameters among nonadherent patients with chronic psychosis. Information about smoking status and antipsychotic nonadherence was obtained via auto-anamnestic and hetero-anamnestic information. PANSS data were obtained while patients were in a psychotic state during the illness requiring hospitalization. Plasma total cholesterol, LDL cholesterol, HDL cholesterol (HDL- c), triglyceride, and glucose levels were determined after a 12-hour fasting period. Compared with non-smoking antipsychotic-naïve first-episode individuals, antipsychotic-naïve smokers exhibited significantly lower depression factor scores, and significantly higher triglyceride levels and triglyceride/HDL- c ratio (p < 0.05). Compared with non-smoking nonadherent chronic individuals, nonadherent smokers exhibited significantly lower negative symptoms and negative factor scores, and lower HDL- c levels. Contributions of smoking to clinical and metabolic parameters ranged from ~ 3.4 % to 10 %. Our present results indicated that smoking may be associated with less severe clinical psychopathology, and with increased risk for metabolic abnormalities, among unmedicated patients with first-episode psychosis and chronic psychosis.