本研究根據Phillips, Wu and Yu (2011)以及後續相關文獻所提出的檢測泡沫模型對台灣市場以及NASDAQ指數進行實證研究。本文使用的模型分別為PWY模型、PSY模型、Rolling Window ADF，以及我們參考PSY模型與Rolling Window ADF所建構出的Rolling Window BSADF。我們利用上述四種模型對NASDAQ指數進行泡沫檢測，以及在台灣上市公司股票中建構投資泡沫投資組合與不投資泡沫投資組合。實證結果顯示投資泡沫投資組合績效優於不投資泡沫投資組合，此結果與Guenster et al. (2009)相同，同時本研究所建構的Rolling Window BSADF在投資績效上優於另外三種模型；此外對NASDAQ指數的檢測發現Rolling Window BSADF 具有檢定結果獨立於起始點的選取與不受週期性泡沫破裂影響等優點，故綜合以上實證結果，Rolling Window BSADF 對於泡沫的檢測與建構泡沫投資組合的績效明顯優於另外三種模型。 / This paper used the bubble examination model according to Phillips, Wu and Yu (2011) and following papers to conduct empirical research on Taiwan market and NASDAQ index. The models used in this paper are PWY model, PSY model, Rolling Window ADF and Rolling Window BSADF that referred to the PSY model and the Rolling Window ADF. We tested NASDAQ index through the above models to test the bubbles, and constructed the portfolio of investing bubbles against not investing. The result shows that the portfolio of investing bubbles performs better than not investing bubbles, which is the same as the result of Guenster et al. (2009). In addition, the Rolling Window BSADF constructed by this paper are superior to the other three models on the performance of investment. Moreover, the examination of NASDAQ index finds that there are some advantages of Rolling Window BSADF including that the test result is independent of the selection of the initial point and not affected by the broken of cyclical bubbles and so on. To sum up, this paper concludes that the bubble examination and the construction of bubble investing portfolio of the Rolling Window BSADF are significantly better than the other three models.