An analytical setup for respiration rate measurements was developed and evaluated in pelagic water samples using a commercially available optical oxygen sensor (Optode™). This setup required the development of a gas tight stopper to connect the sensors to a 1 dm 3 glass sample bottle, precise temperature control (± 0.05°C), and proper stirring of samples. The detection limit and precision of the method was 0.3 mmol O 2 m -3 d -1 . This was similar to the detection limit for the high-precision Winkler titration method reported in field studies. When compared with the Winkler method, the Optode sensor enabled operator-independent, high temporal resolution measurement of respiration, better coverage of plankton groups and detection of non-linear oxygen decline, without the need for wet chemistry. Respiration rates measured by the Optodes showed good accuracy when compared with measurements made with the Winkler titration method (3% deviation), followed the expected temperature response (Q 10 = 3.0), were correlated with chlorophyll a and were congruent with earlier reported values in the literature. The main source of uncertainty was a necessary correction for system drift during the incubation period, due to oxygen release from the plastic components. Additionally, less stringent temperature control on board research vessels during rough seas reduced the precision. We conclude that the developed Optode system can be used to measure respiration in productive coastal waters. Samples from cold or deep waters were, however, often below the detection limit.