Spinal muscular atrophy (SMA) is a motor neuron disease. Nusinersen, a splice-switching antisense oligonucleotide (ASO), was the first approved drug to treat SMA. Based on prior preclinical studies, both 2′-O-methoxyethyl (MOE) with a phosphorothioate backbone and morpholino with a phosphorodiamidate backbone—with the same or extended target sequence as nusinersen—displayed efficient rescue of SMA mouse models. Here, we compared the therapeutic efficacy of these two modification chemistries in rescue of a severe mouse model using ASO10-29—a 2-nt longer version of nusinersen—via subcutaneous injection. Although both chemistries efficiently corrected SMN2 splicing in various tissues, restored motor function and improved the integrity of neuromuscular junctions, MOE-modified ASO10-29 (MOE10-29) was more efficacious than morpholino-modified ASO10-29 (PMO10-29) at the same molar dose, as seen by longer survival, greater body-weight gain and better preservation of motor neurons. Time-course analysis revealed that MOE10-29 had more persistent effects than PMO10-29. On the other hand, PMO10-29 appears to more readily cross an immature blood-brain barrier following systemic administration, showing more robust initial effects on SMN2 exon 7 inclusion, but less persistence in the central nervous system. We conclude that both modifications can be effective as splice-switching ASOs in the context of SMA and potentially other diseases, and discuss the advantages and disadvantages of each.