In this paper, we use the "Family Resemblance Approach" (FRA) as a framework to characterize how scientists view the nature of science (NOS). FRA presents NOS as a "system" that includes clusters or categories of ideas about the cognitive-epistemic and social-institutional aspects of science. For example, the cognitive-epistemic aspects include aims and values such as objectivity and scientific methods such as hypothesis testing. Social-institutional aspects refer to a range of components including social values such as honesty about evidence and institutional contexts of science such as research institutions. Characterized as such, NOS is thus a system of interacting components. The initial account of FRA was proposed by philosophers of science and subsequently adapted and extended for science education including through empirical studies. Yet, there is little understanding of the extent to which FRA coheres with scientists' own depictions about NOS. Hence, an empirical study was conducted with scientists to investigate their views about FRA as well as their views of NOS using the FRA framework. In so doing, the research sought to explore the utility of FRA from scientists' point of view. Qualitative and quantitative analysis of 17 Taiwanese scientists' responses to a set of written questions indicates that scientists are in agreement with the FRA account of NOS, and they detail all aspects in their reference to NOS, although the social-institutional aspects are underrepresented in their depiction. Implications for further studies and science education are discussed.