Numerous previous studies have shown the effectiveness of music therapy in enhancing cognitive functions in patients with dementia. Despite this, robust evidence in this field, especially concerning the comparison of different music therapy types, is lacking. Therefore, randomized controlled trials (RCTs) focusing on music therapy and cognitive functions in dementia patients, termed by "music" AND "dementia" OR "Alzheimer's disease" AND "cognitive", were identified from primary electronic databases to conduct this network meta-analysis (NMA). The primary outcome focused on the impact on cognitive functions, and the secondary outcome was the comparison of dropout rates between the intervention groups and the usual care control groups. Standardized mean difference (SMD) values and the corresponding 95% confidence intervals (CIs) were computed for effect evaluation. This study protocol has been registered in IPLASY (INPLASY202430082). A total of 14 RCTs with 1056 participants were enrolled, examining interventions including Active Music Therapy (AMT), Active Music Therapy with Singing (AMT + Sing), Rhythmic Music Therapy (RMT), Listening to Music (LtM), and Singing (Sing). The results indicated that RMT, AMT + Sing, and AMT all significantly improve cognitive functions in dementia patients, of which the SMD were 0.76 (95% CI = 0.32-1.21), 0.79 (95% CI = 0.03-1.49), and 0.57 (0.18-0.96), respectively. Compared with the control group (usual care), no music therapy type was associated with an increased dropout risk. In conclusion, music therapy can improve cognitive functions in patients with dementia without increasing the risk of dropout, particularly RMT, AMT + Sing, and AMT.