The purposes of this research are to study structure of Thailand foreignexchange market and to test its efficiency . Thailand foreign exchange market is structurally!!y small because governmentonly allows some businesses involving with foreign currency to buy They areimporter,exporter,comPany and business that have foreign currency assets orliabilities,foreign investor with baht account, and commercial bank. Commercial banksare also al!owed to dealer. For this ARCH effect was used to determine the efficiency of Thailand foreignexchange market to see the effect of time varyingIt volatility . Then, theCointegration and Error Correction Model was tested for the market efficiency . Thisresearch used daily data and weekly data of spot and forward exchange rate of usdollar , Japanese yen and Deutsche mark against baht.The data was divided into twotime periods. The first period was from the period of basket exchange rate regimebetween -1 July 1996 to 1 July 1997 The second period was from the period of managedfloat exchange rate regime between 2 July 1997 to 2 July 1998. It can be concluded from the results that in the long run Thailand foreignexchange market is not efficient in al! cases but the in short run Thailand foreignexchange market is efficient on!y during the period of managed float exchange rateregime of B:100Y, This analysis includes the risk premium which was neglected Ink theprevious model. The fact that inefficiency of Thai!and foreign exchange market can beexplained by the risk premium and irrational! expectation of investors. the policiesregarding reduced risk premium and rational expectation of investor in the foreignlexchange market,such as policies for prevention for volatile excahage agereconstructive organization,and transparency for disseminated dats, are recommendedas means of increasing the market efficiency in the future.