As technology improves, the number of teachers (at schools, universities or in companies as well) using Web Based Training (WBT) is growing. Every day they have new and enhanced ways to meet their students' needs. The nature is an immense source of inspiration to explain and to solve complex problems in WBT (individualization, optimization, etc). Unfortunately, Nature-Inspired Modeling (NIM), when used in an innovation of WBT productions, imposes new requirements and constraints. It requires great thought and concentration, takes much time and demands a multitude of skills, in particular technical skills. The challenge is that the teacher is an educator and a domain expert first; but not trained in required technical skills (web authoring, media creation, etc.). Hence, on getting things done in this case, semantic and didactic features of NIM risk to be diluted during WBT development due to the focus of the teacher on technical tasks and losing focus elsewhere. The purpose of this study is to present a methodology and a cognitively inspired environment (called Moulinette) that was conceived and implemented to support adequately innovative production of WBTs, particularly those based on NIM. In this work, we focus on two perspectives. First, we claim a collaborative production as a way to meet teachers' skills for an optimized WBT production. Secondly, we point out the need of new process that imitates the teachers' thinking from a macroscopic view. As proof of concept, we have used our environment to produce and evaluate collaboratively two WBTs based on NIMs. The capabilities of our approach are compared to the existing ones. The realized experimentations have validated our assumptions and demonstrated that the proposed framework and its design process provide an enhanced assistance during collaborative production of WBTs. Finally, we have demonstrated the benefit of optimizing the production of WBTs using nature-inspired paradigms in.