Nitrous oxide (N2O) emissions from various combustion systems was measured by use of two analytical methods, the performance of which was evaluated together with the sampling procedures. One analytical method is grab sampling followed by analysis by gaschromatography (GC). This procedure was found to give accurate results, provided SO2 and H2O were removed during sampling in order to avoid N2O formation in the sample. An evaluation of a continuous infrared N2O analyzer shows that it is a useful instrument if measures are taken to eliminate or adjust for interferences from CH4, SO2, H2O, and to a minor extent also from NH3. A considerable formation of N2O was found when a gas containing SO2 and NO was scrubbed in a weak acid solution. The sampling program from a variety of combustion systems revelas that N2O emissions usually are low (0–5 ppm) from the combustion of pulverized coal. The results from a gas turbine, a laboratory gas‐fired fiber burner, and a domestic oil combustor are also in the very low range, whereas emissions from combustion of straw or wood were slightly higher. Even higher emissions were observed from circulating fluidized bed combustion (CFBC), from which the level of emissions increased with decreasing combustion temperature and increasing excess air levels. Sampling inside the CFBC reactor showed that N2O formation is particularly rapid at the transition from reducing to oxidizing conditions at the secondary air inlet. Pilot plant experiments with the selective non‐catalytic reduction of NO by urea or ammonia resulted in a large increase in N2O emissions during the injection of urea, whereas almost no increase was found during the injection of ammonia.