Biodiversity preservation is essential for environmental health, which is becoming a valuable indicator for quality of life for human being. In this sense, automatic and bioinspired intelligent system can provide powerful tools for monitoring, identification or tracking of species. Automatic acoustic analysis is a non-invasive methodology, easy to use and useful even in complex conditions and inaccessible environments. This work is focused on the development of an automatic system for analysis of environmental agents and their interaction and it is based on human hearing abilities and several Machine Learning paradigms. This preliminary system is evaluated over a multi-environmental database.