Microgrids have become a viable option for improving the dependability and resilience of dispersed energy infrastructure. Microgrid loads present unique difficulties for fault categorization and identification. The traditional techniques for fault detection are unable to identify the precise directions and small fault currents in the microgrid. Bidirectional current flows exacerbate the issue and result in erroneous tripping. The correct evaluation of fault direction in a microgrid becomes complex with the presence of inverter-based distribution generation. Machine learning-based Decision Tree fault analysis offers a solution for precise fault detection. By incorporating machine learning into microgrid systems, there's potential to establish a more sustainable and resilient energy source. This research paper employs the decision tree algorithm to detect the fault accurately with high efficiency in microgrid. Based on the findings of this paper, it is noted that decision trees (DT) offers a straightforward implementation compared to other techniques. This implies their robustness across different operating scenarios.