This document outlines a research proposal to develop an Oromo stance classification model using machine learning techniques. The problem is that most existing stance detection research focuses on English and European languages. The objectives are to review literature, collect and prepare Oromo language data from Facebook, build and test machine learning models for classification, and evaluate the best-performing model. The scope is political stance detection in Oromo texts. Limitations include not knowing who wrote texts or their legitimacy. The significance is understanding public opinion to help policymaking. The methodology will use a design science research process, collecting data with Facepager and evaluating the model using accuracy, precision, recall, and F1-score metrics.