Abstract : As highlighted in several recent reviews, there is a need to harmonize the way LCA of perennial crops is conducted (Bessou et al. 2013; Cerutti et al. 2013). In most published LCA on perennial crops, the agricultural production is based on data sets for just one productive year. This may be misleading since performances and impacts of the system may greatly vary year by year and the evolution of the stand over the cycle induces specific mechanisms (nutrient re-mobilization, yield alternating, resistance etc.) that must be included. Without a proper mechanistic model, the only way to account for such phenomena is to wider the data sets to at least account for each stage of the stand development and, if possible, all years of the crop cycle. Three modeling choices for the perennial crop cycle were tested in parallel in two contrasted LCA case studies: oil palm fruits from Indonesia, and small citrus from Morocco. Modeling choices tested were: i) a chronological modeling over the complete crop cycle of orchards (Bessou et al. 2013), ii) a three years average from the productive phase and iii) a selection of different single years from the productive phase. In both case studies, the system boundary included all processes from the seed production until the harvested fruits at farm-gate. The functional unit was 1 kg of fresh fruits. The chosen approach to model the perennial cycle influences the final results and deserves specific attention.