The results of student mathematics learning in Indonesia are still relatively low, both at national and international levels. This is because in the learning process the teacher still uses theapproach teacher-centered. So that students are not optimal in developing their knowledge. Therefore, one learning model that can observe this needs to be pursued in a form of learning that is able to increase student activity and learning outcomes is one of thelearning models Guided Discovery Learning. The objectives of this study are (1) To compare the learning outcomes of students taught withmodels Guided Discovery Learning with student learning outcomes taught with conventional learning, (2) To find out the improvement in student learning outcomes taught withlearning models Guided Discovery Learning and ( 3) To find out the response of students after taking part in learning with thelearning model Guided Discovery Learning for students at Bandar Baru Middle School. The research method used was a type of Quasi Experiment with the type of control group pre-test post-test design. The population in this study were all eighth grade students of Bandar Baru Junior High School andsamples were taken, Random namely class VIII3 as the experimental class and VIII6 as the control class. Data collection uses learning outcomes tests, and student response questionnaires. Analysis of student learning outcomes data used t-test and N-gain while analysis of student response data was carried out through analysis of average score criteria. From the results of data processing it can be concluded that the learning outcomes learned usinglearning models Guided Discovery Learning are better than student learning outcomes taught with conventional learning,learning models Guided Discovery Learning can improve student learning outcomes and student responses to learning models Guided Discovery Learning very positive.