PURPOSE: The purposes of this study were to develop and test a prediction model for caregiving experiences including caregiving satisfaction and burden in dementia family caregivers. METHODS: The stress process model and a two factor model were used as the conceptual frameworks. Secondary data analysis was done with 320 family caregivers who were selected from the Seoul Dementia Management Survey (2014) data set. In the hypothesis model, the exogenous variable was patient symptomatology which included cognitive impairment, behavioral problems, dependency in activity of daily living and in instrumental activity of daily living. Endogenous variables were caregiver's perception of dementia patient's unmet needs, caregiving satisfaction and caregiving burden. Data were analysed using SPSS/WINdows and AMOS program. RESULTS: Caregiving burden was explained by patient symptomatology and caregiving satisfaction indicating significant direct effects and significant indirect effect from unmet needs. The proposed model explained 37.8% of the variance. Caregiving satisfaction was explained by patient symptomatology and unmet needs. Mediating effect of unmet needs was significant in the relationship between patient symptomatology and caregiving satisfaction. CONCLUSION: Results indicate that interventions focusing on relieving caregiving burden and enhancing caregiver satisfaction should be provided to caregivers with high levels of dementia patients' unmet needs and low level of caregiving satisfaction.