A critical step in authentication and security is online signature verification. In recent years, signature verification methods have significantly improved thanks to the quick development of deep learning. An overview of deep learning-based online signature verification is provided in this work. In this paper, we introduce a time- aligned recurrent neural network (TARNN)-based method for online signature verification. By matching the signature image in time and feeding it to the TARNN model, the suggested approach captures both static and dynamic aspects of the signature. A fully connected layer for classification follows a bidirectional LSTM layer in the TARNN model. The system is tested against common benchmark datasets after being trained on a sizable dataset of real and fake signatures. The suggested system achieves state-of-the-art performance in online signature verification, according to experimental results. The technology can be utilized for trustworthy and effective signature verification in a variety of applications, including banking, security, and online shopping.