This paper examines the efficiency of intraday price discovery in the Seoul won/dollar FX market. More specifically, we investigate three aspects of the efficiency of price discovery: the speed of price discovery, the degree of price discovery per trade, and the accuracy of price discovery in the market. The speed and degree of intraday price discovery are measured by WPC (weighted price contribution) and WPCT (WPC per trade), which are devised by Barclay and Warner (1993) and Barclay and Hendershott (2003), respectively. The accuracy of intraday price discovery is measured by VR (variance ratio) proposed by Lo and MacKinlay (1988). We use the real- time transaction data and two-minute-interval quotes data of inter-dealer trades, brokered through Seoul Money Brokerage Services Ltd. from April 1 to May 30, 2003. Price discovery is defined as the dynamic process by which prices incorporate new information in financial markets including FX market. Since price discovery is generally the most important function in financial markets, its efficiency has become a very valuable public good. Despite the importance of its economic implications, the studies on the efficiency of intraday price discovery on FX market are very limited. Previous studies can be generally classified into two categories: those that belong to the first category indirectly analyze the efficiency of intraday price discovery by comparing it with that of other FX markets, while those of the second category directly analyze an aspect of the efficiency of intraday price discovery on an FX market. Rosenberg and Traub (2006) and D’Souza (2007) fall into the first category. Rosenberg and Traub (2006) examines the relative information shares of spot and futures FX markets using the methodology of Hasbrouck (1995) for two sample periods: May-August 1996 and March-May 2006. D’Souza (2007) investigates information content of order flows in inter-dealer brokered markets during two years from October 2000 to September 2002 using real-time transaction data. He reports that trades are more informative when they are initiated in a local country rather than in major foreign exchange centers like London and New York. Kaul and Sapp (2009) is the only existing study that falls into the second category. The study analyzes the accuracy of inter-dealer Euro-USD and CAD-USD FX markets for the year of 2000 through the VR test and GARCH analysis, and reports that the relative accuracy of price discovery is the greatest when trading activity is high and dealer concentration reaches its peak. The distinct features of our paper from the previous domestic and international studies on the efficiency of intraday price discovery are summarized as follows. First, this is domestically the first paper to directly examine the efficiency of intraday price discovery. Our paper provides another angle for the second category in the field, which currently includes the only paper by Kaul and Sapp (2009). Second, as far as we know, this is domestically and internationally the first paper to study all the dimensions of the efficiency of price discovery. Kaul and Sapp (2009), the only existing study of the second category, on the other hand, has a limitation of examining only accuracy out of three dimensions of efficiency of price discovery. Third, this paper provides an answer to the following question: “To what degree is the efficiency of intraday price discovery in the Seoul FX market attained during a trading day?” The answer is yet to be provided by the existing domestic literature. Eom et al. (2008) finds that the autocorrelation of daily won-dollar exchange rates disappeared after January 2001. This result indicates that the efficiency of price discovery has been achieved in the market at least in the frequency of daily since then. There has not been any further study which investigates the degree of intraday efficiency of price discovery is attained. The overall results of this study can be summarized as follows. First, the intraday pattern of price discovery exhibits an inverse J-shape. In addition, most of intraday price discovery is attained during the first 90 minutes just after the market opening. To be more specific, during the close-to-close (open-to-close) time span, about 75.3% (44.5%) of price discovery occurs during the first 90 minutes just after the market opening. After that, the speed and degree of price discovery decrease up to the last 60 minutes before the market closing, from which they increase again. Second, the efficiency of intraday exchange rates, after the market opening, is attained up to lunch break. During the period from the market opening to lunch break, there exists the autocorrelation in won/dollar exchange rates. When the short horizon is expanded from 30 minute to 120 minute with the long horizon fixed at a trading day, the computed VRs are significantly greater than 1. This result indicates that won/dollar exchange rates under-react to information in the short horizon, and thus there exists a positive autocorrelation in exchange rate returns. Meanwhile, if the short horizon is expanded to 150 minute i.e. the entire morning session, then the computed VR is not significantly greater than 1, indicating that positive autocorrelation does not exist. Overall, our findings imply that the efficiency of price discovery in the Seoul won/dollar FX market is very high; the process of price discovery occurs very quickly with great accuracy.