Transportation is an important tool and carrier for people to realize their trip purposes. Thus, it's a vital measurement for studying spatio-temporal pattern of individuals. Trip chain refers that a series of displacements completed by an individual in order to do one or more activities using a transportation. The time period of trip chain is one day. There are lots of information on individuals' trip purpose contained in trip chains. Extract this information from trip chains help to explore individuals travel behavior, which help understand the urban space. In previous studies, researchers have been focused on inferring and exploring the dynamic characteristics of commuting behavior, go to school activities, go home activities, entertainment, and leisure activities in urban space with the help of smart card data and taxi trajectory data. But limited studies have been carried on detecting hospital-seeking behavior with the assistance of trip record. With this in mind, this paper attempted to extend the application of trip records on hospital-seeking behavior. Specifically, we proposed a theoretical framework used to detecting hospital-seeking behavior from trip record. It consists of six principle, such as proximity principle, ring-closure principle, single-purpose principle, infrequency principle, time-coherence principle, and accompany principle. Also, a methodology for detecting hospital-seeking behavior was put forward based on the theoretical framework. Taking Beijing as an example, we found the key parameters of detecting hospital-seeking behavior and calibrated their thresholds. Finally, spatial and temporal patterns of hospital-seeking behavior were explored to reveal the accuracy of our results. On the one hand, the spatial patterns of hospital-seeking behavior showed that patients were mainly concentrated in tertiary hospitals. Tertiary hospitals have better professional skills and a larger service area than secondary and primary hospitals. Thus, they attracted and served more patients. On the other hand, patients' arrival time shows a high peak during 8:00 am and 10:00 am and a low peak during 13:00 pm and 15:00 pm, which closed to start time of registration and treatment. Two aspects above both supported the accuracy of results and rationality of the theoretical framework. The application of trip chains on detecting hospital-seeking behavior could make up for the shortages of traditional data, which is a small sample and difficult to access. This paper provided a new perspective, methodology, and data source for researching hospital-seeking behavior. Moreover, it could provide methodology references for other activities based on trip records.