This study examined the influence of microwave (MW) power (300–800 W) and standard geometries (slab, cube, disc, sphere) on banana drying kinetics. Nine thin-layer models were explored. Model validation employed the coefficient of determination and root mean square errors, further checked using residual sum of squares and zero mean of errors. The Henderson and Pabis model emerged as optimal for thin-layer drying. As MW power ranged from 300 to 800 W, drying times reduced: slabs (700–220 sec), cubes (800–300 sec), discs (560–160 sec), and spheres (620–200 sec). Moisture diffusivities spanned 9.12×10−9 to 7.2×10−6 m2/s, revealing reduced moisture diffusion at lower MW power. Activation energies were tabulated for discs (25.54 W/g), spheres (108.96 W/g), cubes (88.41 W/g), and slabs (107.56 W/g). Sample mass uncertainty was approximated at 0.03 g (∼1% error). Specific energy varied between 140 and 244 kJ, while microwave energy values ranged from 44.58 to 112.8 MJ/kg, declining with increased power. Cube samples showed maximum energy consumption, whereas disc samples typically consumed the least. Notably, disc-shaped bananas demonstrated peak drying energy efficiencies (41.75–51.3 %) across all MW power levels. Average drying efficiencies were observed between 19.97 and 51.3 % for the studied power range.