Hasil studi pendahuluan memperlihatkan bahwa kegiatan belajar mengajar fisika tidak ada interaksi dua arah antara guru dan siswa serta jarang melakukan kegiatan praktikum. Selain itu, keterampilan menarik kesimpulan siswa juga sangat rendah. Penelitian ini difokuskan pada upaya peningkatan keterampilan menarik kesimpulan siswa melalui penerapan interaksi kelas co-constructive berbasis investigasi sains. Tujuan penelitian ini adalah untuk mengetahui peningkatan keterampilan menarik kesimpulan sebagai hasil dari penerapan interaksi kelas co-constructive berbasis investigasi sains serta mengetahui proses yang terjadi selama interaksi kelas co-constructive berbasis investigasi sains dilaksanakan. Metode penelitian yang digunakan adalah kuasi eksperimen dengan desain one group pretest-posttest design. Sampel penelitian ini adalah 40 orang siswa kelas X di salah satu SMA Negeri di Kota Bandung. Pengumpulan data dilakukan melalui tes dan lembar observasi. Diperoleh hasil bahwa keterampilan menarik kesimpulan siswa mengalami peningkatan dengan skor n-gain sebesar 0,41 yang termasuk dalam kategori sedang. Adapun proses yang terjadi selama interaksi kelas co-constructive dan investigasi sains dilaksanakan, secara kuantitatif investigasi sains lebih dominan, akan tetapi jika ditinjau secara kualitatif interaksi kelas co-constructivelah yang berperan dalam mempengaruhi peningkatan keterampilan menarik kesimpulan siswa SMA.The results of a preliminary study show that the learning and teaching physics has no two-way interaction between teacher and student and has a few practical activities. In addition, students’ capability to draw conclusions is also very low. The research focused on improving the skills of students through the application of drawing conclusions through co-constructive classroom interactions which is based on science investigation. The purpose of this study was to determine the enhancement of ability of drawing conclusions as a result of the application of co-constructive classroom interactions based on scientific investigation and to know the processes that occur during the interaction of co-constructive classroom-based science investigation carried out. The method used is a quasy-experimental design with pretest-posttest one group design. This study sample was 40 students in class X in one of the Senior High school in Bandung. The data was collected through testing and observation sheet. The results infer that the students’ skills in drawing conclusion were increased by a score of n-gain of 0.41 is included in the medium category. From quantitative perspective, the processes that occur during the interaction of constructive and co-class scientific investigation carried out of science is more dominant, but when it is viewed by qualitative one, co-constructive classroom interaction play a role in influencing the improvement the skills of senior high school students to draw conclusions.