Identification of Hand-written digits is a rational key point in pattern identification applications. There are many uses of hand-written digits identification like mail sorting in postal, cheques processing in the banks, data entry through forms, etc. The key to the issue lies in the expertise to grow a wellorganized algorithm that can accept hand-written numbers and which are submitted by end-users by the scanners, tablets, and other digital devices. This paper gives a viewpoint to handwritten numbers recognition constructed on machine learning models, and deep learning models and shows the outcomes in the shape of accuracy. The primary objective of this paper is to guarantee powerful and dependable methodologies for the acknowledgment of handwritten numbers using machine learning and deep learning algorithms. Several machine learning algorithms such as Decision Tree (DT), Naïve Bayesian (NB) classifier, Multilayer Perceptron (MLP), Support Vector Machine (SVM), Random Forest (RF), and deep learning algorithms such as Convolutional Neural Network (CNN), AlexNet, and Multilayer Perceptron (MLP) have been used for recognition of hand-written digits in Jupyter Notebook and Matlab.Through some features extraction, and different experiments and analysis of Machine Learning Algorithms (MLA) and Deep Learning Algorithms (DLA), the accuracy of deep learning algorithms is better than the machine learning algorithms.