Smart wellness services collect various types of lifelogs such as walking steps and sleep duration via smart devices. However, most of the existing smart wellness services focus on displaying each individual lifelog to users through graphs and charts. Therefore, they have limitations on supporting overall and easy understanding of various lifelogs. A lifelogs-based daily wellness score (LDWS) is a useful tool to resolve such limitations. LDWS combines the lifelogs into a score to represent an overall level of daily health behaviors, thus supporting overall and easy health behavior monitoring of users. This research developed LDWS as part of developing a smart wellness service for college students (SWSCS) in collaboration with an IT company. Lifelogs of 41 college students were collected through a four-week pilot run of SWSCS and were subsequently fitted to a random effects model. Based on the model estimates, LDWS was determined by linearly aggregating seven behavior variables. The utility of the developed LDWS was validated through a second pilot run of SWSCS. This paper also discusses the potential use of LDWS for SWSCS and the factors to be considered in the development of a lifelogs-based wellness score for a smart wellness service. This research is expected to contribute to advancing smart wellness services with lifelogs.