Agroecological studies on sugarcane dealing with genotype by environment by management interactions commonly generate complex datasets. Data are collected separately and can be multi-scale (from phytomer to field), multi-species (sugarcane and companion plants), and multi-disciplinary (agronomy, entomology, ecophysiology, weed science, etc), making their use complex in system data analysis. This paper presents a solution to manage these heterogeneous data using database and metadata technology. A relational database, named ECOFI, was designed from the analysis of the content and the structure datasets of multi-disciplinary experiments with sugarcane. The results of this analysis showed that most datasets shared data corresponding to common environmental by management factors and the same measurements, such as yield, weed occurrence and insect incidence. However, each of them had its own structure. To analyse biology by environment by management interactions efficiently, the structure of ECOFI was built into three parts: the environmental conditions, the agricultural practices, and the impacts observed on biotic and abiotic variables using agronomic measurements. In standard databases, each additional observed variable generally forces an update to the existing database model. The model of the ECOFI database does not require such modification. Taking into account a new variable is very easy, consisting solely of the addition of a new record in a table, after having stored the new variable label and its definition (unit, type, scale) in a generic metadata table. This technology minimizes the number of tables, columns and empty cells and improves database query performance. ECOFI is already used to manage, explore and analyse the data from agroecological experiments, ecophysiological observations and sugarcane modelling in Réunion Island and other sugar-producing countries. ECOFI is a generic database that improves analysis and facilitates access to multi-disciplinary