Science education in elementary schools plays a vital role in developing scientific literacy, problem-solving skills, and reflective awareness. However, classroom practices are often teacher-centered and rely on printed worksheets that limit student engagement and competency development. This study aimed to design and evaluate a STEM-based Electronic Student Worksheet (E-LKPD) integrated with Deep Learning principles; joyful, meaningful, and mindful learning to enhance students’ science competence in IPAS learning, which integrates natural and social sciences to support interdisciplinary understanding and real-world problem solving. The study employed a Research and Development (R&D) approach adapted from Borg and Gall, consisting of six stages: needs analysis, product design, expert validation, limited trial, revision, and large-scale implementation. Participants were 32 elementary students from a suburban public school with diverse academic and socioeconomic backgrounds. Data were collected using expert validation sheets, student response questionnaires, observation forms, and pretest–posttest assessments. Data analysis included descriptive statistics, Content Validity Index (CVI), paired-sample t-tests, and effect size calculations. Results indicated that the developed E-LKPD was feasible, with CVI scores in the valid to highly valid category. Students’ science competence significantly improved, with mean scores increasing from 56.8 to 78.9 (t (31) = 9.21, p < 0.001, d = 1.63). Additionally, more than 75% of students reported positive learning experiences related to joyful, meaningful, and mindful learning. In conclusion, integrating STEM and Deep Learning principles into electronic worksheets is an effective strategy for strengthening elementary students’ science competence and supporting 21st century learning.