<![CDATA[Learning results of physics at SMPN 1st Natar is not maximum because of the conventional teaching method of the teachers. This research compared the learning results of Exclusive model with direct instruction (DI) model, the purpose was to find out the differences between learning results of Exclusive model with DI model and to investigate which models that has higher average. The population was all students of VIII class at SMPN 1 Natar and the samples were students of VIII E and VIII F with Purposive Sampling method. The research design was Pre-Experimental Design with One Group Pretest-Posttest Design type. The hypothesis was analized by using normalize test and Independent Sample t-test, the results of Independent Sample t-test indicated that t-value cognitive result was 3.35, affective was 4.26, and psychomotor was 9.47 higher than t-table (2.04), it means that there were differences of average learning result of Exclusive model with DI model, and the average of Exclusive model was higher than DI model.Hasil belajar Fisika siswa di SMPN 1 Natar belum maksimal disebabkan cara mengajar guru yang konvensional. Penelitian ini membandingkan hasil belajar model Exclusive dengan model Direct instruction (DI), bertujuan untuk mengetahui perbedaan hasil belajar model Exclusive dengan model DI dan manakah yang rata-rata hasil belajarnya lebih tinggi. Populasi penelitian ini yaitu seluruh siswa kelas VIII dan sampelnya yaitu siswa kelas VIII E dan VIII F dengan metode Purposive Sampling. Disain penelitian berupa Pre-Eksperimental Design dengan tipe One Group Pretest-Posttest Design. Hipotesis dianalisis menggunakan uji normalitas dan uji Independent Sample t-test, diperoleh nilai t-hitung kognitif (3.35), afektif (4.26), dan psikomotor (9.47) lebih besar dari t-tabel (2.04), artinya ada perbedaan rata-rata hasil belajar model Exclusive dengan model DI serta rata-rata hasil belajar siswa pada kelas Exclusive lebih tinggi dibandingkan dengan kelas DI.Kata kunci: model exclusive, model direct instruction, hasil belajar]]>
