Microwave diagnostics have been widely utilized to measure the important parameters of high temperature and high density plasmas. Reflectometry is known as a promising microwave diagnostic which has a number of merits to measure electron density profiles. In the KSTAR device, X-mode FM reflectometry is planned to measure the plasma density profiles. FM reflectometry is required to extract phase information on raw mixer IF signals, thereby obtaining time-of-flight of reflectometry signals. It is known that the data analysis method is crucial to determine the performance of FM reflectometry In fact, there are several analysis programs which have been utilized in various FM systems. Since each program was developed for a specific device, however, it is difficult to directly apply it to a different reactor like the KSTAR device. It is necessary, therefore, to develop a data analysis program for the KSTAR FM reflectometry. In this paper, complex digital demodulation (CDM) and wavelet transformation are examined in terms of the performance of density profile reconstruction. For the comparison of both methods, FM reflectometry signals are generated on the basis of assumed profiles and the interaction of the X-mode wave and the plasma. In order to see how well both methods work under various conditions, three types of profiles are assumed and noise effects are included. As a result, both methods work well under the condition of gentle density gradient and small noise level. As density gradient becomes steeper and noise level gets higher. the reconstruction performance of wavelet is better than that of CDM.