In recent years the oxidative potential (OP) emerged as a promising indicator of potential health effects of PM2.5 because it represents the ability of PM to carry or to catalyse reactive oxygen species (ROS) able to produce oxidative stress. Seasonal and spatial trends of natural and anthropogenic sources contributing to PM2.5 and its OP is investigated. The work is based on a dataset collected in 2024 simultaneously at an urban background and a traffic site, in south Italy during cold and warm seasons. A detailed dataset was obtained: elemental analysis (ED-XRF), total and soluble carbonaceous aerosol (OC/EC, WSOC), main ions, organic tracers, and water-soluble organic nitrogen (WSON). Oxidative potential (OP) was measured with ascorbic acid (OPAA) and dithiothreitol (OPDTT) assays. The Positive Matrix Factorization (PMF5) model resolved the contributions of ten distinct sources: vehicular exhaust, biomass burning, resuspension/construction, primary biogenic (fungal spores), nitrate, sulphate, sea spray, aged marine, secondary organic aerosol (SOA), and long-range dust transport. Vehicle emissions predominated for OPAA across both locations, whereas OPDTT was driven by a broader combination of factors, notably traffic, SOA, biomass burning, and dust resuspension. Distinct patterns were observed; OPDTT levels peaked in winter while OPAA in summer. Traffic contributes for 17-30% to OPDTT (25-37% to OPAA) significantly more than the contributions PM2.5 (14-20%). An opposite trend is observed for sulphate contributing for 9-11% to OPDTT and 12-13% to OPAA compared to contributions of 22% to PM2.5 mass at both sites. Secondary aerosols contribute to OPDTTV (26-38%) and to OPAAV (19-21%). These findings suggest implementing seasonally tailored strategies to minimise PM2.5-associated health risks like targeting biomass burning during cold period (22-33% of OPDTT and 15-27% of OPAA) and local soil resuspension during the warm period (19-31% of OPDTT).