<![CDATA[The dominance of conventional teacher-centered instructional models and surface learning orientation systemically results in the stagnation of students' motivation and science learning outcomes in regional schools. This predictive-associative ex-post facto quantitative study aims to analyze the predictive contribution and degree of association of the deep learning model—integrating mindful, meaningful, and joyful learning (3M) pillars—towards students' motivation and cognitive science achievement. Utilizing structured data collection techniques via psychometric questionnaires and objective written tests distributed to a sample of 94 eighth-grade students at SMPN 2 Garawangi, data were inferentially analyzed using simple linear regression modeling and Pearson Product Moment correlation coefficients. The results demonstrated that the deep learning model has a statistically significant positive effect on both student motivation and science academic achievement. Correlation analysis confirmed a significant positive relationship between deep learning implementation and improvements in students' motivation and cognitive achievements. Practically, the implementation of this approach effectively transformed the classroom atmosphere into a more interactive environment and substantially increased students' average test scores beyond the minimum mastery criteria. In conclusion, the deep learning instructional model based on three core pillars effectively serves as a valid statistical predictor to optimize students' academic potential and motivation in junior high school science education. ABSTRAK Dominasi model instruksional konvensional yang berpusat pada guru (teacher-centered) dan berorientasi hafalan (surface learning) secara sistemik berimplikasi pada stagnasi motivasi serta hasil belajar sains siswa di daerah. Penelitian kuantitatif berdesain ex-post facto asosiatif-prediktif ini bertujuan untuk menganalisis kontribusi prediktif dan derajat asosiasi model pembelajaran deep learning yang mengintegrasikan pilar mindful, meaningful, dan joyful learning (3M) terhadap motivasi dan prestasi belajar kognitif IPA siswa. Melalui teknik pengumpulan data terstruktur memanfaatkan kuesioner psikometrik dan tes tertulis objektif yang didistribusikan kepada sampel sebanyak 94 siswa kelas VIII di SMPN 2 Garawangi, seluruh data dianalisis secara inferensial menggunakan pemodelan regresi linear sederhana dan koefisien korelasi Pearson Product Moment. Hasil penelitian menunjukkan bahwa model deep learning terbukti memberikan pengaruh positif dan signifikan secara statistik terhadap motivasi belajar maupun prestasi belajar IPA siswa. Analisis korelasi mengonfirmasi adanya hubungan positif yang signifikan antara penerapan deep learning dengan peningkatan motivasi dan prestasi belajar kognitif peserta didik. Secara praktis, implementasi pendekatan ini terbukti mentransformasi atmosfer kelas menjadi lebih interaktif dan meningkatkan capaian rata-rata ulangan harian siswa secara nyata melampaui Kriteria Ketuntasan Minimal. Simpulan penelitian menegaskan bahwa pemodelan deep learning berbasis tiga pilar utama efektif bertindak sebagai prediktor statistik yang valid untuk mengoptimalkan potensi akademik dan motivasi sains siswa di lingkungan sekolah menengah pertama. ]]>
