In Laser Powder-Directed Energy Deposition, creating Single Scan Tracks is an approach to find the optimal process parameters, reducing cost and time. However, the geometric evaluation and categorization of good Single Scan Tracks is a time-consuming manual task. This approach employs machine learning algorithms to estimate the geometrical quality of the cross section derived from Single Scan Tracks. A dataset was compiled from 194 cross-sections of 49 Single Scan tracks deposited with AISI 316L powder steel. The laser power and powder feed rate were designated input parameters, While the height-width ratio and dilution were considered output parameters. Support Vector Machine, Stochastic Gradient Descent, and Random Forest were employed to classify optimal beads and enhance the decision-making process. The authors validated the efficacy of the proposed methodology by utilizing a classification report obtained for each machine-learning algorithm. The accuracies obtained from the Support Vector Machine, Stochastic Gradient Descent, and Random Forest were 93%, 92%, and 92%, respectively. Support Vector Machine achieved the best performance in recognition due to its ability to predict the highest quality of beads and qualities beyond the evaluation range.