This study sought to determine the level of utilization of portfolio assessment and the level of numerical competence of the at risk junior high school students. Eventually, it sought to determine the significant relationship between the independent and the dependent variables. This study made use of a descriptive correlation non-experimental quantitative research method. It was descriptive because the data were presented in quantitative descriptions on the “Portfolio Assessment and Numerical Competence of at-risk Junior High School Students”. According to Gay (2006) descriptive-correlation design determines and supports the way things are. Using a questionnaire in conducting a survey, interview or observations usually collects the data. The study was conducted in Digos City National High School, Division of Digos City. The respondents of this study were the 200 at risk junior high school students of the research locale who answered the questionnaire. Universal sampling procedure was utilized in this study in determining the respondents considering the large number of respondents in the research locale. This study revealed that in terms of basic competencies in mathematics, at risk junior high school students displayed good performance in the basics of mathematics. This finding was the basis for policy recommendation that would address the weak points of the at risk junior high school students. The school needed to design a program that would address the grey areas of the results of the study.