This dataset presents a reconstructed and standardised compilation of dissolved trace metal (dTM) concentrations across global marine environments. It integrates observations from peer-reviewed literature and publicly available marine datasets to enable consistent, cross-study comparison of trace metal distributions in the ocean. The dataset includes six key trace metals: Copper (Cu), Cadmium (Cd), Zinc (Zn), Manganese (Mn), Iron (Fe), and Lead (Pb), all of which play critical roles in marine biogeochemical cycling and ecosystem functioning. The compilation is based on 82 peer-reviewed studies spanning multiple ocean basins. In addition, publicly accessible datasets were incorporated, while GEOTRACES Atlas visualisations were used as a qualitative benchmark to support interpretation of concentration ranges in open-ocean settings. These visualisations were not used as direct measurements but served as contextual validation for reconstructed values. All dissolved trace metal concentrations were harmonised to nanomolar (nM) to ensure consistency across studies reporting in diverse units and analytical frameworks. The dataset is structured to support large-scale synthesis by aggregating observations according to geographic location, marine environment classification (open, coastal, shelf, offshore, and inshore systems), and standardised water depth categories (0–100 m, 101–1000 m, ≥2000 m). For each unique combination of location, marine environment, and depth category, concentration statistics were derived, including minimum, maximum, and mean values. Where multiple observations were reported within a grouping, ranges were preserved to capture variability, while mean values were calculated for comparative analysis. In cases where only summary statistics (mean ± standard deviation) were available in source material, the mean ± standard deviation was calculated. Observations digitised from figures were processed using consistent extraction and standardisation procedures. A structured quality control workflow was applied throughout dataset reconstruction. Records flagged in original studies as contaminated, unreliable, or analytically unsuitable were excluded. Observations below detection limits were omitted from range calculations where explicitly indicated. Additional filtering ensured that only records with sufficient spatial, environmental, and methodological metadata were retained for inclusion. Marine environment classifications were assigned based on original study descriptions or inferred from sampling context, bathymetry, and proximity to continental margins where not explicitly provided. This harmonisation enables consistent comparison across heterogeneous sampling designs and reporting formats. The dataset also includes a qualitative GEOTRACES-based comparison flag, indicating whether reconstructed concentrations fall within expected ranges for open-ocean environments as depicted in the GEOTRACES Atlas. This field is intended solely as a contextual reference and does not represent direct GEOTRACES measurements. This dataset is suitable for applications in marine biogeochemistry, including inter-basin comparisons, statistical and spatial analysis of trace metal distributions, and ecosystem or Earth system modelling. It also supports investigations of trace metal cycling, anthropogenic influences, and natural variability across marine environments. Users should note that inherent variability exists due to differences in sampling strategies, analytical methods, temporal coverage, and spatial resolution across source studies. Some values were reconstructed from digitised figures, which introduces additional uncertainty. Consequently, this dataset represents a harmonised, derived product and should be used alongside original source data for full methodological transparency. This reconstructed dataset is a processed derivative of a broader raw extraction database and is intended to support reproducible, large-scale synthesis of dissolved trace metal distributions in the global ocean.