Cross comparison of satellite-derived land surface phenology (LSP) and ground measurements is useful to ensure the relevance of detected seasonal vegetation change to the underlying biophysical processes. While standard 16-day and 250-m Moderate Resolution Imaging Spectroradiometer (MODIS) vegetation index (VI)-based springtime LSP has been evaluated in previous studies, it remains unclear whether LSP with enhanced temporal and spatial resolutions can capture additional details of ground phenology. In this paper, we compared LSP derived from 500-m daily MODIS and 30-m MODIS-Landsat fused VI data with landscape phenology (LP) in a northern U.S. mixed forest. LP was previously developed from intensively observed deciduous and coniferous tree phenology using an upscaling approach. Results showed that daily MODIS-based LSP consistently estimated greenup onset dates at the study area (625 m × 625 m) level with 4.48 days of mean absolute error (MAE), slightly better than that of using 16-day standard VI (4.63 days MAE). For the observed study areas, the time series with increased number of observations confirmed that post-bud burst deciduous tree phenology contributes the most to vegetation reflectance change. Moreover, fused VI time series demonstrated closer correspondences with LP at the community level (0.1-20 ha) than using MODIS alone at the study area level (390 ha). The fused LSP captured greenup onset dates for respective forest communities of varied sizes and compositions with four days of the overall MAE. This study supports further use of spatiotemporally enhanced LSP for more precise phenological monitoring.