Microfluidic devices have presented some advantages when compared to conventional processes. The major advantages include the use of reduced amounts of reagents and samples, shorter reaction times, lower manufacturing cost, higher surface area versus volume, higher heat / mass transfer, and greater control over field strengths and concentration of molecules. These devices have been used in a variety of applications including, but not limited to, chemical and biological sensors and chemical processes. To identify molecules and/or compounds, the components are usually identified by techniques characterized as off-line, with the need for sample withdrawals, which in many cases do not allow real-time monitoring of the process. Some alternative techniques such as Raman spectroscopy allow the real-time monitoring of chemical processes being a viable alternative to monitor and quantify chemical species. Thus, this paper evaluates the use of Raman spectroscopy to monitor and quantify the concentration of ethanol in a PDMS (Polydimethylsiloxane) microfluidic device in two different mixtures of compounds: water/ethanol (for control and predictive ability of the spectrometer) and vegetable oil/ethanol in an online way. Spectra of the blends were obtained in glass beaker (batch) and during the flow process in the microdevice (in situ). Through multivariate analysis of the data, PLS (partial least squares) models were constructed. It was possible to demonstrate the on-line detection of ethanol concentration in water and vegetable oil in PDMS microfluidic device with excellent coefficients of determination (> 0.9) and with an analysis time of 45 s.