This research focuses on using advanced pattern recognition techniques to identify characters and extract information from ancient inscriptions. It makes use of artificial neural networks such as Multilayer Perceptron (MLP) and Feedforward Neural Networks (FNNs) to address the challenges posed by deteriorating inscriptions. The process consists of training neural networks, extracting features from images, and preprocessing images. While FNN is a good starting point for evaluation, MLP excels at detecting complex patterns, resulting in higher recognition accuracy. The study also improves character recognition performance by combining neural networks and pattern recognition algorithms.