The problem in this research is whether the Problem-Based Learning model can improve the ability to write compound sentences in Grade IV of SDN No. 94 Kota Utara, Gorontalo City. This study aims to improve the ability to write compound sentences through the Problem-Based Learning model in Grade IV of SDN No. 94 Kota Utara, Gorontalo City. This research is a Classroom Action Research (CAR). The data collection techniques used in this research include written tests, observation, and documentation. Initial observation results showed that out of 15 students, only 1 student (7%) was categorized as proficient, 4 students (27%) as less proficient, and 10 students (67%) as not yet proficient. In Cycle I, Meeting 1, improvement occurred with 2 students (13%) categorized as proficient, 5 students (33%) as less proficient, and 8 students (53%) still not proficient. More significant progress was seen in Cycle I, Meeting 2, with 10 students (67%) categorized as proficient, 3 students (20%) as less proficient, and only 2 students (13%) still not proficient. In Cycle II, the final results showed that 13 students (87%) were categorized as proficient, 2 students (13%) as less proficient, and no students remained in the not proficient category. ABSTRAKMasalah dalam penelitian ini adalah apakah melalui model pembelajaran Problem Based Learning kemampuan menulis Kalimat Majemuk Di Kelas IV SDN No. 94 Kota Utara Kota Gorontalo dapat meningkat. Penelitian ini bertujuan untuk meningkatkan kemampuan menulis Kalimat Majemuk melalui model pembelajaran Problem Based Learning Di Kelas IV SDN No. 94 Kota Utara Kota Gorontalo. Jenis penelitian ini adalah penelitian Tindakan kelas (PTK). Teknik pengumpulan data dalam penelitian ini meliputi tes tertulis, observasi, dan dokumentasi. Hasil observasi awal menunjukkan bahwa dari 15 siswa, hanya 1 siswa (7%) tergolong mampu, 4 siswa (27%) kurang mampu, dan 10 siswa (67%) belum mampu. Pada siklus I pertemuan 1, peningkatan terjadi dengan 2 siswa (13%) tergolong mampu, 5 siswa (33%) kurang mampu, dan 8 siswa (53%) masih tidak mampu. Kemajuan terlihat lebih signifikan pada siklus I pertemuan 2, dengan 10 siswa (67%) tergolong mampu, 3 siswa (20%) kurang mampu, dan hanya 2 siswa (13%) masih belum mampu. Pada siklus II, hasil akhir menunjukkan bahwa 13 siswa (87%) sudah tergolong mampu, 2 siswa (13%) kurang mampu, dan tidak ada lagi siswa yang tergolong tidak mampu.