Comparison of single trait of selection  (total green forage yield) via combining ability test, with multiple trait selection (total green forage yield, dry weight of root nodules and seed yield ) by using independent  culling levels or Index selection for the improvement of barseen clover (Trifolium alexandrinum, L.) were obtained from 100 polycrosses isolated from a base population represented the second generation of random mating for a seed synthetic composed from 23 farmer's seed lots of Meskawi (Multi-cut) type. Selection intensity was 10% for the three systems. Selection for multiple trait was significantly much rewarding than single trait selection. A realized gains of 12.20, 17.40, 17.20, 28.50, 14.10, 17.90 and 8.55% from index selection were obtained vs.6.58, 5.96,3.79, 10.27,7.87,55.05 and 4.47% from single trait recurrent selection for total green forage , dry forage , protein and seed yields, seed index, dry weight of nodules and leaves/stem ratio, relative to the base population. Although index selection gave higher magnitudes of realized gain relative to either the base population or the check varieties, these figures were not significantly different from those obtained with independent culling levels. The amount of effort required for index selection method is somewhat greater beside that, breeder should wait until all observations are recorded to construct an index. These are not essential with independent culling.