The goal of collecting data resulting from activities of manufacturing workshops is to facilitate the workshop control and to feed all decision levels with adequate, updated and reliable information. On each level, decision making needs information relevant to: time, quality and cost. So corresponding data must be identified and stored, then processed and provided adequately and efficiently for each decision activity. Looking for this goal, we propose a global approach for manufacturing data organisation and a methodology for implementing this organisation on a large variety of production systems. The approach is based on events and states identification of main manufacturing entities, then on their storage in an adequate data base structure. The approach interests are underlined and practical examples are given.