In-situ elemental analysis by laser-induced breakdown spectroscopy (LIBS) for future lunar missions requires comprehensive investigation due to the lack of atmosphere and the presence of surface materials with diverse chemical compositions (from anorthosites to basalts) as well as physical properties (from fine regolith to boulders). We developed and cross-validated a multivariate LIBS calibration model by measuring 169 compressed fine powders of geologic samples, fully encompassing the bulk composition range of lunar meteorites, in vacuum. We further validate this model by measuring lunar meteorites, terrestrial anorthites, and lunar simulants in various physical forms, including rock chips and soils with varying grain sizes and bulk density. Quantification accuracy, evaluated using root mean squared error (RMSE), yielded 2.5 wt% SiO2, 0.25 wt% TiO2, 1.2 wt% Al2O3, 1.3 wt% MgO, 1.2 wt% CaO, 0.33 wt% Na2O, 0.47 wt% K2O (0.060 wt% K2O in the <1 wt% range), and 1.5 wt% T-Fe2O3 for fine powders. For rock chips, the RMSEs were 3.1 wt% SiO2, 0.32 wt% TiO2, 2.2 wt% Al2O3, 2.5 wt% MgO, 2.0 wt% CaO, 0.33 wt% Na2O, 0.089 wt% K2O, and 2.1 wt% T-Fe2O3. Despite significant difference in physical conditions between powders and chips, their RMSE remained consistent within a factor of two. The grain size and bulk densities of soils showed minor effects, as their errors mostly fell within the RMSEs of rock chips. These RMSEs verify that the quantification accuracy of LIBS is sufficient for distinguishing the subgroups within the lunar anorthosite suite (e.g., anorthosites vs. norites) and basalts (e.g., high-Ti vs. low-Ti) for coarse-to-fine soils, loose-to-compact soils, and rocks. Furthermore, we show that LIBS can differentiate between “purest” and “pure” anorthosites (98 and 95 vol% plagioclase anorthosites) based on the 3σ detection limits of Mg and Fe lines. These capabilities of LIBS align well with the goals of future lunar exploration, such as locating ilmenite-rich soils for resource extraction, purest anorthosites for understanding the early lunar evolution, and noritic impact melts for improving the chronology. Our results demonstrate that LIBS serves as a versatile tool for rapid geochemical characterization on the Moon.