This paper proposes an innovative approach for generating the topology of artificial neural networks. The proposed method utilizes an iterative function system (IFS) to produce coordinate coordinates for neurons, enabling the construction of the input layer, hidden layer, and output layer topologies. The suggested method differs from conventional fully connected or convolutional neural networks in that it does not depend on a predetermined architecture. Instead, it is built dynamically through iterative calculations. The experiment involved the construction of a feedforward neural network with a single hidden layer. The connection topology of the neural network was generated using the IFS approach. Empirical findings demonstrate that the produced neural network has a high level of responsiveness to sine wave input signals. Additionally, an assessment was conducted on the network’s performance, encompassing the computation of the correlation coefficient between the output signal and the input signal, the variance and standard deviation of the output signal, as well as the reaction time. This study presents a novel approach for the automated development of neural network architecture, demonstrating great potential.