This study employs the K-Nearest Neighbors (KNN) algorithm to classify households into poor and non-poor groups based on data derived from the National Household Consumption and Expenditure Survey 2013/2014, a decennial initiative led by the High Commissioner for Planning in Morocco. Utilizing a non-parametric, instance-based approach, the KNN algorithm determines predictions by assessing the similarity between the target instance and its neighboring examples. Through training the KNN model on a dataset containing relevant demographic and socioeconomic traits, effective categorization of families, particularly in identifying those living in poverty, is achievable. This research aims to assess the efficiency of the KNN algorithm in household classification and its potential to provide valuable insights into poverty alleviation efforts, reevaluating its performance using reformulated metrics. The presented results demonstrate the model’s high accuracy (0.98) and F1 Score (0.99), indicating a robust capability to accurately classify households. The precision (0.73) and specificity (0.73) metrics reveal the model’s ability to discriminate non-poor households, However, the sensitivity (0.99) indicates how well it can identify low-income households.