Bilge water is created when ship tanks are flushed and engine rooms are cleaned with water and detergents/degreaser. The resulting oil containing bilge water is usually treated in an on-ship waste water-treatment unit. To facilitate minimization of water-treatment chemicals and to assure that the oil concentration is low enough to allow discharge in the sea, it is important to know the oil concentration in the outflow. In this work we have tested and developed passive acoustic measurements for measurement of oil concentration in bilge water. The starting point was to make synthetic bilge water to test the concept and to get information about the system, including the properties of oils used and the specifics of applying acoustic measurement to them. We investigated the effect of salt in the bilge water by modelling synthetic bilge water with and without NaCl. The prediction errors obtained was 2.5 ppm oil (samples with salt), 0.6 ppm oil (samples without salt) and 4.5 ppm oil (samples with and without salt) (0.5-18 ppm calibration range). Thereafter real samples were collected from a ship. Since the samples had either high or low oil concentration, they were mixed in order to get a representative set of calibration samples with a large range of concentrations. The oil concentrations in the samples were determined by laboratory analysis, the experiments were performed and the calibration models developed. The best model was validated using a separate test set and yielded a test set prediction error of 7.3 ppm. This can be regarded as a good result, since the average absolute error for the laboratory measurement is around 5 ppm, and acoustic measurements is a promising technology to determine oil in bilge water. Recommendations for further work is to lower the error of the laboratory reference method used to determine the oil concentration, to improve the sampling handling chain and possibly also to improve the acoustic measurement equipment before new trials are performed.