The limitations on spatial resolution and on the availability and measurement accuracy of remote sensing images are the primary problems in the estimation of the large-scale planting area for maize. The integration of mid- and low-resolution images is the one of primary methods used for the estimation of large-scale crop planting areas using remote sensing. The use of a single-temporal thematic mapper (TM) image results in a low accuracy of maize recognition, so a mid-scale time-series normalized difference vegetation index (NDVI) dataset, which was derived from the fusion of the moderate-resolution imaging spectroradiometer (MODIS) and TM images based on the wavelet transform, was established. The planting area was estimated using the minimum distance model and the accuracy was evaluated using in-situ samples. The results show that the estimation of the maize-sown area based on the time-series NDVI information of the integrated images reached high levels of gross and position accuracy (89% and 90%), indicating that this method could fully utilize the time-series information from the MODIS images and the spatial resolution of a TM image. The use of the difference in phenophases among fall crops enables the effective classification of the spatial distribution of these crops.