Background: This study aimed to evaluate changes in compound muscle action potential (CMAP) amplitude in adults with spinal muscular atrophy (SMA) undergoing nusinersen treatment and its association with motor function improvements. Methods: This multicenter study assessed median, ulnar, and peroneal CMAP over a follow-up of up to 4.5 years using linear mixed models. Motor function was measured using the Revised Upper Limb Module (RULM) and the Hammersmith Functional Motor Scale Expanded (HFMSE). Correlations between CMAP and motor function scores were analyzed. Results: Seventy-eight patients (27 ambulatory, 51 non-ambulatory) were included. Baseline ulnar CMAP ≥ 2.0 mV distinguished SMA type 3 from type 2 with 91.3% sensitivity and 88.9% specificity (AUC 0.96, 95% CI 0.92–1.0), while baseline median nerve CMAP ≥ 6.5 mV showed 91.7% sensitivity and 77.3% specificity (AUC 0.84, 95% CI 0.72–0.96). No significant changes over time were observed in median, ulnar, and peroneal CMAP amplitudes (p > 0.05). CMAP trajectories did not differ between SMA types 2 and 3 (p > 0.05). No significant difference in the change in RULM or HFMSE at any time point was observed between SMA patients with baseline median nerve CMAP < 5 mV and those with CMAP of ≥ 5 mV (p > 0.05). No significant correlations were found between changes in median nerve CMAP and HFMSE or RULM (p > 0.05). Discussion: CMAP amplitudes remained stable during nusinersen treatment, with no differences in trajectories between SMA types 2 and 3. Our findings suggest that while CMAP amplitude correlates with disease severity, it may not serve as a sensitive biomarker of treatment response in adult SMA patients.