Problem-based learning (PBL) is an approach that improves students' skills in problem solving. The application of PBL as an approach of teaching in a class requires students' presence; such constraint cannot be fulfilled by all students. Therefore, it is important to avoid this problem by implementing an online PBL environment where students are grouped remotely and work together to solve a problem proposed by the teacher. This will guide the learning process of the learners and can evaluate their solution. In reality, we can find members who do not really contribute to solve a problem. From this point of view, the teacher's evaluation will not be adequate to estimate the contribution of the learner in the solution of a given problem. Therefore, it is important to think of another way for assessing learners' solution. So, the challenge is to implement an online PBL environment and to propose a new method for assessing students. In this paper, the authors present their system called Problearn. The developed system allows students to solve problems remotely in small groups. Furthermore, the system evaluates each student based on his behavioral profiles during the problem-solving process. To do so, the system must keep track of different actions carried out by the students. This system has been tested by students of a computer science department where they achieved very good results.