Here we present a method for in situ determination of stable antimony (Sb) isotope compositions by ultraviolet (UV)-femtosecond-laser-ablation-multi-collector-ICP-MS (fs-LA-MC-ICP-MS). Metallic antimony and a number of Sb minerals (stibnite, senarmontite, chalcostibite, tetrahedrite, boulangerite, bournonite, zinkenite, pyrargyrite and dyscrasite) have been investigated. In order to verify the results of the in situ method, LA-MC-ICP-MS measurements were compared with solution MC-ICP-MS analyses on two chemically homogeneous stibnite samples and Sb metal. The internal precision of in situ measurements was better than 0.045 parts per thousand. The long-term reproducibility for these three materials was better than 0.1 parts per thousand. These results imply that any of these three materials may be used as an in situ Sb isotope standard. All delta Sb-123 values were determined by applying a standard-sample bracketing protocol and mass bias correction using Sn (NIST SRM 3161a standard solution) isotope ratios. Since no certified Sb isotope standard is currently available, in situ isotope analyses were performed in bracketing to a stibnite in-house standard and subsequently recalculated relative to NIST SRM 3102a. Matrix effects from Cu, Pb, Ag, Fe and Zn on the natural mineral Sb isotope ratios were insignificant. The mass interference of Te-123 on Sb-123 can only precisely be corrected for materials with Te/Sb <= 0.2. The LA and solution analyses of delta Sb-123 homogeneous zinkenite, dyscrasite and pyrargyrite agree excellently with each other. Senarmontite shows more variable delta Sb-123 at small scales, determined with LA-MC-ICP-MS, but homogeneous delta Sb-123 determined by solution analyses with a small offset between solution and LA analyses. Other minerals, like boulangerite, show heterogeneous delta Sb-123 (approximate to 0.5 parts per thousand) for both solution and LA analyses. Overall, these results demonstrate that fs-LA-MC-ICP-MS is suitable to measure Sb isotopic ratios in Sb-rich sulfides, sulfosalts and oxides (excluding Sb tellurides) with a precision better than 0.1 parts per thousand and a spatial resolution of approximate to 30 mu m. Thus, it may be used as a tool to analyse spatially resolved Sb isotope composition at the mineral or even sub-mineral scale, e.g. to address the processes of ore formation or Sb redistribution in near-surface environments.