Working memory (WM) function can be reduced in patients suffering from unilateral hearing loss (UHL) and can affect their academic performance. We aimed to compare the WM abilities of three categories of children with UHL: patients implanted with hearing aids (HAs), patients receiving a bone-anchored hearing implant (BAHI), and subjects who did not receive hearing devices. A randomized clinical study, in which 45 children (mean age: 9.5 years) were evaluated by pure tone audiometry (to identify the side and the severity of the UHL), was conducted in a tertiary referral center. Patients were simply randomized into three groups: (1) children without HAs (No-HA group), (2) patients with a (digital) HA (HA group), and (3) children with a BAHI (BAHI group). Their working and short-term memories were studied in both noisy and silent conditions at the recruiting time (T0, baseline) and 6 months after (T1) the treatment. Statistical analyses were performed to analyze the variances between T0 and T1 within each group and between the three groups. The No-HA group improved its T1 WM scores in silence (p < 0.01), but not in noise. The HA and BAHI groups showed statistically significant variances of T1 WM in noise (p < 0.01 and p < 0.01, respectively). The HA and BAHI groups did not show statistically significant variances compared to T1. Our results suggest that hearing devices (HA and BAHI) in children with sensorineural UHL (SUHL) can improve WM capacity in noise. We speculate that bilateral hearing capacity might improve the quality of life of this population, especially during everyday activities where noise is present.