The quality of model outputs and advices are tributary of the data quality. Providing data quality requires first to analyze methods of data collection, data storage and data extraction, second to propose methods of data validation and finally to develop metrics of data quality. Such an approach of data quality is based on principles of quality management system like those defined by ISO9001:2000 international standard. The MEQUAPRO project developed within the framework of program SIDEPECHE at the Ifremer, aims at reaching this requirement. Data qualification consists in assigning a quality value to a data according to a range of preset quality values. Qualification is based on a validation by one or more methods (primarily the respect of protocols), which enables to check if a data reaches the preset quality value. The following stage consists in passing from qualification to quality indicators. These indicators make possible to check and follow in time if data and products (estimates from data) achieved a precise qualitative aim, to meet either internal needs of the Ifremer or those for "external customers" (French Ministry for Agriculture and Fishing, European Union, etc). The French process of data collection in board constrained by the European Data Collection Regulation illustrates this approach.