The inclusion of nature of science (NOS) in science curricula has long been advocated as a key means to accomplish scientific literacy. For example, in the U.S., the Science for All Americans (AAAS, 1989) and Benchmarks for Science Literacy (AAAS, 1993) included an explicit emphasis on the teaching of NOS. More recently, A Framework for K-12 Science Education (NRC, 2012) also highlighted that 'there is a strong consensus about characteristics of the scientific enterprise that should be understood by an educated citizen' (p. 78). While these and other science curriculum documents around the world have emphasised the inclusion of NOS since the 1990s, there is currently a lack of empirical evidence on teachers' responses to curriculum reforms focused on NOS and particularly on their views of and practices in the classroom assessment of NOS. In an attempt to fill this gap, this study investigates Korean science teachers' practices of classroom assessment in the wake of the recent national curriculum reform that foregrounds NOS as an explicit learning expectation. In the 2015 National Curriculum of Korea, a new subject named 'scientific inquiry and experimentation' (SIE) was introduced as a separate compulsory subject and is being taught to Grade 10 students. SIE features NOS, which has epistemic characteristics distinguished from those of scientific content knowledge, as one of its main learning goals. To understand how teachers approach the assessment of this NOS-centred curriculum, I focus on one school where three teachers taught and assessed NOS through a number of different assessment tasks. Interviews, classroom observations, assessment materials and various lesson artefacts collected over two years are used to illustrate how the three teachers viewed and practised the classroom assessment of NOS. Several important aspects of the teachers' views and practices relating to the classroom assessment of NOS are reported. First, I discuss the mismatch that existed between teachers' NOS teaching intentions and NOS assessment intentions, and between their views and enactments of the summative assessment of NOS. The three teachers enacted the same set of NOS summative assessment activities in different ways, and their enactments were associated with their perceived purposes and foci of NOS assessment. Second, the teachers' collaborative decision making about NOS assessment revealed several challenges of NOS summative assessment, such as distinguishing 'understandings' and 'views' of NOS, ensuring fairness among different enactments of assessment tasks, and controlling the literacy and content knowledge demand when assessing NOS. Third, an in-depth analysis of one teacher's NOS instruction suggested that her feedback and other formative assessment practices mediated the NOS learning goal of the lesson with science content knowledge, history of science and other aspects of NOS, and built connections between these elements. Throughout the lesson, the teacher created NOS learning opportunities by eliciting, responding to and acting upon evidence of student learning. Given the centrality of high-quality classroom assessment in student learning, the findings will be useful to inform future curriculum and assessment reforms as well as to better support teachers for teaching and assessing NOS. Based on the existing literature and the findings of this study, I assert that supporting teachers for meaningful and effective classroom assessment of NOS is key to empowering students to develop an elaborate understanding of what science is and how it works.