Both northern and southern Sudan are deploying artemisinin-based combinations against uncomplicated Plasmodium falciparum malaria (artesunate+sulfadoxine-pyrimethamine [AS+SP] in the north, artesunate+amodiaquine [AS+AQ] in the south). In 2003, we tested the efficacy of 3 day AS+SP and AS+AQ regimens in vivo in the isolated, seasonally endemic Nuba Mountains region (the first study of AS combinations in southern Sudan). We also analysed pre-treatment blood samples for mutations at the P. falciparum chloroquine transporter (Pfcrt) gene (associated with CQ resistance), and at the dihydrofolate reductase (Dhfr) gene (associated with pyrimethamine resistance). Among 161 randomized children under 5 years, PCR-corrected cure rates after 28 days were 91.2% (52/57, 95% CI 80.7-97.1) for AS+SP and 92.7% (51/55, 95% CI 82.4-98.0) for AS+AQ, with equally rapid parasite and fever clearance. The Pfcrt K76T mutation occurred in 90.0% (144/160) of infections, suggesting CQ would work poorly in this region. Overall, 82.5% (132/160) carried mutations at Dhfr (N51I, C59R or S108N, but not I164L), but triple mutants (more predictive of in vivo SP failure) were rare (3.1%). CQ use should be rapidly discontinued in this region. SP resistance may propagate rapidly, and AS+AQ is likely to be a better long-term option, provided AQ use is limited to the combination.