Effective population size (Ne) is an important concept to understand the evolution of a population. In conservation, Ne is used to assess the threat status of a population, evaluate its genetic viability in the future and set conservation priorities. An accurate estimation of Ne is thus essential. The main objective of this thesis was to better understand how the estimation of Ne using molecular markers can be improved for use in conservation genetics. As a first step, we undertook a simulation study where three different methods to estimate Ne were investigated. We explored how well these three methods performed under different scenarios. This study showed that all three methods performed better when the number of unlinked loci used to make the estimation increased and the minimum number of loci need for an accurate estimation of Ne was 100 SNPs. A general assumption in the estimation of Ne is that genetic drift is homogeneous throughout the genome. We explored the variability of Ne throughout the genome of the Danish Holstein cattle, using temporally-spaced samples of individuals genotyped with a 54K SNP chip. We found heterogeneity in Ne across the genome both between chromosomes and in genomic windows within chromosomes. Heterogeneity in Ne has implications for conservation management as Ne is used to evaluate the threat status of populations. Ne can vary locally along the genome, hence a population can be wrongly classified if heterogeneity in Ne is not taken into account when assessing the population against threat status thresholds. When molecular markers are not available, populations can be managed using pedigree information. However, this is challenging to do so for group-living species since individuals and their parentage are difficult to determine. We adapted a pedigree-based method and developed bioinformatics tools to manage this type of species. The results obtained in this thesis have contributed to a clearer understanding of: the limits of methods to estimate Ne, the evolutionary processes affecting genetic drift in the genome at a local scale, and finally the development of a method to manage group-living species. These results provide additional knowledge to improve the management of conservation programs, that overall, will contribute to the conservation of genetic biodiversity.