This dataset contains radiomic features extracted from structural T1-weighted MRI scans of participants diagnosed with First-Episode Psychosis (FEP) and matched Healthy Controls (HC). A total of 304 subjects were included, comprising 130 FEP and 174 HC cases. Preprocessing and cortical/subcortical segmentation were performed with FreeSurfer v7.4.2, and radiomic features were extracted using PyRadiomics v3.0.1 following IBSI guidelines. The dataset includes: Subject identifiers (anonymized) Diagnostic labels (FEP vs HC) Extracted radiomic features across cortical and subcortical regions ( aparc+aseg Freesurfer in first-order, shape, and multiple texture families: GLCM, GLRLM, GLSZM, NGTDM, GLDM). This resource can support reproducibility and benchmarking in neuroimaging-based machine learning studies, enabling the exploration of structural MRI biomarkers in psychosis. This work was supported by the Valencian Regional Government (CIPROM/2023/28) and the Carlos III Healthcare Institute (ISCIII), projects PI20/00473 (to M.D. Moltó), PI22/01985 (to O. Rivero) and CP21/00059 (to O. Rivero), co-funded by the European Union.