In this study, temporal MODIS data are used to detect rice planting date in Taiwan. However, the time-series satellite data are easily contaminated by noises such as cloud cover and atmospheric conditions. Empirical Mode Decomposition (EMD) and wavelet transform are applied for data noise filtering, and the planting date is specified by detecting the local maximum and inflection points from the smoothed temporal NDVI profile. The results indicate that the time-series data filtered by EMD are useful for detecting rice planting dates. Moreover, the detecting results obtained from EMD are more accurate than from wavelet transform.