This study aimed to determine the ripeness of bananas using hyperspectral surface analysis and how a rapid UV-C (ultraviolet-C light) surface treatment could reduce decay. The surface of the banana fruit and its stages of maturity were studied using a hyperspectral imaging technique in the visible and near infrared (370–1000 nm) regions. The vselected color ratios from these spectral images were used for classifying the whole banana into immature, ripe, half-ripe and overripe stages. By using a BP neural network, models based on the wavelengths were developed to predict quality attributes. The mean discrimination rate was 98.17%. The surface of the fresh bananas was treated with UV-C at dosages from 15–55 μW/cm2. The visual qualities with or without UV-C treatment were compared using the image, the chromatic aberration test, the firmness test and the area of black spot on the banana skin. The results showed that high dosages of UV-C damaged the banana skin, while low dosages were more efficient at delaying changes in the relative brightness of the skin. The maximum UV-C treatment dose for satisfactory banana preservation was between 21 and 24 μW/cm2. These results could help to improve the visual quality of bananas and to classify their ripeness more easily.