The target of breeding red-fleshed apples is to increase their potential health benefits related to red flesh coloration and consumer acceptance. The objective of this study was to determine the usefulness of four clones (90, 120, 156, and 158) of red-fleshed apples for freeze-drying compared to the cultivar ‘Trinity’. Red-fleshed apples were dried in the form of slices using a laboratory freeze-dryer. The changes in color features and image texture parameters after drying and the sensory quality of freeze-dried samples were assessed. Trends of increase in the value of the L* parameter and decrease in the a* and b* parameters after freeze-drying were observed. Furthermore, freeze-drying caused statistically significant changes in analyzed image textures named XHMean, RHMean, SHMean, VHMean, LHMean, and UHMean. Machine learning models developed based on the color parameters L*, a*, and b* distinguished raw and freeze-dried red-fleshed apples with an average accuracy of 84% for clone 90 up to 99.0% for clone 156, and models based on twenty selected image textures exhibited an accuracy of 98.5% for clone 156 to 100% for clones 90 and 158 and the cultivar ‘Trinity’. The very attractive external appearance, medium-intense fruity smell, crunchiness, and intense fruity taste of all the apple slices were revealed. The innovative aspect of this study included the comparison of the drying behavior and sensory quality of the new clones and a standard cultivar of red-fleshed apples. Moreover, innovative methods and results were used to determine the effect of freeze-drying on red-fleshed apple quality, considering novel models involving thousands of image textures and machine learning algorithms.