The objective of this investigation was to study the effect of two different methods of selection (modified mass and family selection) for crown diameter on forage yield and quality in Ismailia 91 alfalfa variety (Medicago sativa L.). In 2007/2008, two hundred plants (5% intensity of selection) were selected for crown diameter, using the modified mass selection method. Equal seeds from each selected plant were bulked to form each of two selection cycles. Plants selected for C1 were also raised as half-sib families in 2008/2009 and selection was practiced between and within half-sib families for the best 10 families (5% intensity of selection). Seeds of selected half-sib families and both modified mass selection cycles C1 and C2 along with the base population were evaluated for forage yield and quality in 2009/2010 season. The realized gains after the two mass selection were 9.77, 12.68, 14.94, 14.00, 11.34,-14.91 and 12.47% for crown diameter, root length, fresh forage yield, dry forage yield, crude protein, crude fiber and ash (%), respectively over the base population. Gains from family selection as% of base population were 21.24, 16.91, 17.24, 16.00, 16.49,-16.41 and 18.90% for these traits in the same order. All studied traits were positively correlated but the correlation between crude fiber (%) and other traits were negative. However, two half sib-families (No. 1 and 6) were significantly higher than the original population for all studied traits except crude fiber (%). Results suggest that both mass and family selection for crown diameter resulted in great improvement of forage yield and quality. In addition, family selection appeared to be more rewarding than mass selection in improving yield and quality of alfalfa.