In the realm of pattern recognition, the automated detection of handwritten text or symbols poses intricate challenges in the field of handwriting recognition. The paper introduces a novel approach that considers the batch_size to calculate the accuracy of the machine learning method to attain precise and effective recognition of handwritten characters. The suggested technique employs artificial neural networks, particularly convolutional neural networks (CNNs), to train a model skilled of precisely identifying and categorizing handwritten characters. Experimental result indicates that the proposed model achieves a notable accuracy percentage of 99.10% on a widely used data set (NIST). The attained high level of accuracy accentuates the effectiveness and viability of machine learning algorithms in research works related to handwriting recognition. The achievement also paves the way for numerous potential applications in fields like document analysis, optical character recognition, and interfaces based on handwriting.