The human brain is a highly connected and integrated system. Local stroke lesions can evoke reorganization in multiple functional networks. However, the temporally-evolving patterns in different functional networks after stroke remain unclear. Here, we aimed to investigate the dynamic evolutionary patterns of functional connectivity density (FCD) and strength (FCS) of the brain after subcortical stroke involving in the motor pathways. Eight male patients with left subcortical infarctions were longitudinally examined at five time points within a year. Voxel-wise FCD analysis was used to identify brain regions with significant dynamic changes. The temporally-evolving patterns in FCD and FCS in these regions were analyzed by a mixed-effects model. Associations between these measures and clinical variables were also explored in stroke patients. Voxel-wise analysis revealed dynamic FCD changes only in the sensorimotor and cognitive regions after stroke. FCD and FCS in the sensorimotor regions decreased initially, as compared to controls, remaining at lower levels for months, and finally returned to normal levels. In contrast, FCD and FCS in the cognitive regions increased initially, remaining at higher levels for months, and finally returned to normal levels. Most of these measures were correlated with patients' motor scores. These findings suggest a network-specific dynamic functional reorganization after stroke. Besides the sensorimotor regions, the spared cognitive regions may also play an important role in stroke recovery.