This study was a pre-experimental study through one group pretest-posttest design. The objective of the study was to see whether there was an increase in generic science skills of students with problem-based learning model in the group activities. The subjects were the chemistry students of 2nd semester in Teacher Training and Education Faculty of Unsyiah. The data collection was conducted by using the instruments such as group activity and multiple choice of generic science skill. The data of group activities was processed from the observation scores while the pretest and posttest were processed by using the average of N-Gain. Based on the data analysis, there were significant differences in the pretest and posttest results KGS after the application of PBL teaching model. Referring to the N-Gain, an increase of KGS occured in all indicators. The highest increase was found in indirect observation while the lowest in logic inference. In indirect observation and cause-effect law, the consistent logic framework gained an increase in the medium category.