CIRAD (www.cirad.fr) and PalmElit SA (www.palmelit.com) are implementing a Marker-Assisted Selection (MAS) of the oil palm through a long-term multi-stage project in order to make optimum use of its existing network of field experiments while capitalising upon recent advances in molecular marker technology. Biotechnology projects, performed in both E. guineensis and E. oleifera species, combine genetics, structural genomics, functional genomics and whole genome sequence information. A QTL (Quantitative Trait Locus) analysis designed for multi-parent populations was designed for crosses between heterozygous parents and tested in oil palm (Elaeis guineensis Jacq.), which is a diploid cross-fertilizing perennial species with small family sizes tested in breeding trials. The algorithm, MCQTL OutBred available for any allogamous species, an extension of MCQTL (INRA, http://carlit.toulouse.inra.fr/MCQTL/), a sofware application to handle multiallelic QTL detection in multicross design, was used to compare two types of QTL search for small size families, within-family analysis and across-family analysis. Results showed that the across-family analysis proved to be efficient due to the interconnected families, but the small family size issue is just partially solved. A conclusion is that a "real-time" MAS of the oil palm perennial crop can be performed, apart a genome wide methodology, by generalising a multi-parent QTL/Gene detection approach through adapted classical genetic trials, novel Field Methodological Systems and Multi-parent QTL/Gene mapping methods and extensive catalogues of expressed Elaeis genes for intra-gene markers, all information analyses being supported by the South Green Bioinformatics Platform (SGBP) applied by CIRAD to the genomic resource analysis of southern and mediterranean plants. Therefore, our presentation focuses on how our current biotechnology researches are integrated into the genetic improvement plan for optimizing the MAS creation of new