We investigate at a novel type of nonlinear discrete-time fractional neural networks using the h-discrete nabla ABC fractional operator. To derive appropriate criteria for the existence and uniqueness of solutions to the issues at hand, we apply basic fixed-point theory techniques. Furthermore, The Ulam-Hyers stability of the considered model, as well as significant findings, are shown. Moreover, to highlight the validity of the presented conclusions, two and three-dimensional examples are explored.