Although most coke oven research is focused on the emission of polycyclic aromatic hydrocarbons, well-known carcinogens, little has been done on the emission of volatile organic compounds, some of which are also thought to be hazardous to workers and the environment. To profile coke oven gas (COG) emissions, we set up an open-path Fourier transform infrared (OP-FTIR) system on top of a battery of coke ovens at a steel mill located in Southern Taiwan and monitored average emissions in a coke processing area for 16.5 hr. Nine COGs were identified, including ammonia, CO, methane, ethane, ethylene, acetylene, propylene, cyclohexane, and O-xylene. Time series plots indicated that the type of pollutants differed over time, suggesting that different emission sources (e.g., coke pushing, quench tower, etc.) were involved at different times over the study period. This observation was confirmed by the low cross-correlation coefficients of the COGs. It was also found that, with the help of meteorological analysis, the data collected by the OP-FTIR system could be analyzed effectively to characterize differences in the location of sources. Although the traditional single-point samplings of emissions involves sampling various sources in a coke processing area at several different times and is a credible profiling of emissions, our findings strongly suggest that they are not nearly as efficient or as cost-effective as the continuous line average method used in this study. This method would make it easier and cheaper for engineers and health risk assessors to identify and to control fugitive volatile organic compound emissions and to improve environmental health.