There is some evidence about the efficacy of preferred music on agitation in elders with dementia; however, little is known about its effectiveness on agitation when implemented by nursing staff in long-term care facilities. Even less is known about use of preferred music for managing anxiety in those with dementia. This quasi-experimental study aimed to evaluate the implementation of a preferred music intervention delivered by nursing staff on agitation and anxiety of institutionalised elders with dementia. The sample comprised of 57 elders with dementia residing in two building complexes which provided similar care routines and staffing in a large Taiwanese residential care facility. These two building complexes were randomly assigned as the experimental and control group. Nursing staff in the experimental group received a facilitation program to prepare them for implementing the preferred music intervention; whereas nursing staff in the control group received no facilitation program. The music intervention based on each resident's music preferences was then provided by the trained nursing staff for 32 experimental residents twice a week for six weeks. Meanwhile, 25 residents in the control group only received the usual standard care without music. All residents were assessed by Cohen-Mansfield Agitation Inventory (CMAI) for overall and three subtypes of agitated behaviours and by Rating of Anxiety in Dementia for anxiety at baseline and week 6. Additionally, the modified CMAI measured the 30-minute occurrence of agitation at baseline, session 4, and session 12. The results indicate that institutionalised elders with dementia who received six weeks of preferred music intervention implemented by trained nursing staff had significant reductions on overall, three subtypes of agitated behaviours, anxiety, and 30-minute occurrence of agitation over time compared to those who received the usual standard care without music. Preferred music shows promise as a strategy for reducing agitation and anxiety in those with dementia when implemented by trained nursing staff. Such intervention can be incorporated into routine activities to improve the quality of care provided by nursing staff and the quality of life of those with dementia in long-term care settings. Our study results provide clinically relevant evidence which contribute to closing the gap between research and practice.