Robot-assisted therapy is regarded as an effective and reliable method for the delivery of highly repetitive training that is needed to trigger neuroplasticity following a stroke. This study focuses on the pre-defined research gaps by combining adaptive assist-as-needed control and immersive virtual environment (VE) into a rehabilitation system, with special attention on the rehabilitation of fine hand motion skills. The virtual reality-based integrated rehabilitation system using Oculus Rift DK2 head mounted display, featuring rehabilitation gaming system with rendering immersive VE is proposed to strengthen activities of daily living and task-oriented kinematic features such as force, range of motion (ROM) and finger coordination. The effectiveness of the system is examined by conducting clinical trials on a group of 8 subacute stroke patients for a period of six weeks with 18 training sessions. The results are verified through clinical evaluation methods and key kinematic features such as active ROM and finger strength. By comparing the pre- and post-training results, the study demonstrates that the proposed method can enhance the efficiency of fine hand motion rehabilitation training by 35% on average in the patients who participated in the experimental work.