Bilateral arm training (BAT) presents as a promising approach in upper extremity (UE) rehabilitation after a stroke as it may facilitate neuroplasticity. However, the effectiveness of BAT is inconclusive, and no systematic reviews and meta-analyses have investigated the impact of different factors on the outcomes of BAT. This systematic review and meta-analysis aimed to (1) compare the effects of bilateral arm training (BAT) with unilateral arm training (UAT) and conventional therapy (CT) on the upper limb (UL) motor impairments and functional performance post-stroke, and (2) investigate the different contributing factors that may influence the success of BAT. A comprehensive literature search was performed in five databases. Randomized control trials (RCTs) that met inclusion criteria were selected and assessed for methodological qualities. Data relating to outcome measures, characteristics of participants (stroke chronicity and severity), and features of intervention (type of BAT and dose) were extracted for meta-analysis. With 25 RCTs meeting the inclusion criteria, BAT demonstrated significantly greater improvements in motor impairments as measured by Fugl-Meyer Assessment of Upper Extremity (FMA-UE) than CT ( = 3.94, = < 0.001), but not in functional performance as measured by the pooled outcomes of Action Research Arm Test (ARAT), Box and Block Test (BBT), and the time component of Motor Function Test (WMFT-time) ( = 0.28, = 0.313). The superior motor impairment effects of BAT were associated with recruiting mildly impaired individuals in the chronic phase of stroke ( = 6.71, < 0.001), and applying a higher dose of intervention ( = 6.52, < 0.001). Subgroup analysis showed that bilateral functional task training (BFTT) improves both motor impairments ( = 7.84, < 0.001) and functional performance ( = 1.02, = 0.049). No significant differences were detected between BAT and UAT for motor impairment ( = -0.90, = 0.681) or functional performance ( = -0.09, = 0.457). Thus, our meta-analysis indicates that BAT may be more beneficial than CT in addressing post-stroke UL motor impairment, particularly in the chronic phase with mild UL paresis. The success of BAT may be dose-dependent, and higher doses of intervention may be required. BFTT appears to be a valuable form of BAT that could be integrated into stroke rehabilitation programs. BAT and UAT are generally equivalent in improving UL motor impairments and functional performance.