Abstract: Studies on method of analysing tar and nicotine in sidestream smoke was conducted using a modified smoking machine equipped with fishtail chimney designed according to the CORESTA recommended method for trapping sidestream smoke, and associating with GC and UV. Two groups of experiments were conducted to evaluate repeatability of the method. The data on sidestream smoke were compared with those obtained by Japanese Tobacco Corporation (JT) during the same period. The results showed that the repeatability of the method was pretty well. The coefficients of variation of all analytical items, such as puff number, tar and nicotine condensed on the wall of the fishtail chimney and trapped by Cambridge filter pad, were less than 10% and those of total tar and total nicotine in sidestream smoke were below 5%. The results were close to the data obtained by JT. It was concluded that the method was suitable for the analysis of tar and nicotine in sidestream smoke. Meanwhile, the principles of sidestream smoke trapping and the selection of analytical conditions were also discussed.