These article shows efficiency of Problem-Based Learning (PBL) in academic performance of course "Physics I", specifically in how learning arises across experience. For this, existing methodologies were adapted in PBL in order to generate six methodological proposals, developed during an academic semester. Impact of PBL was evident at end of semester. It was validated through non-parametric test from Mac Nemar, with a confidence level of 95%. Conclusion is PBL turned out to be highly effective, especially at highest levels of academic performance (demanding application and analysis skills) in which there are found substantial differences with regard of group of control. On the other hand, in area of comprehension, such a difference was not noticed, which indicates that application of above-mentioned methodology (regarding this capability) is not interesting for teacher. This investigation contributes with a program that relies on instruments that measure comprehension, application and analysis of beginning of classic physics in students of course "Physics I". Likewise, a guide of fieldwork provides with guidelines of employment for this course in top level, and examples for its later use.