The risk of post-operative paralysis is high in patients with eloquent area gliomas and without apparent limb motor dysfunction preoperatively. This study aims to evaluate the risk of postoperative paralysis via task-based and resting state fMRI. 24 patients with eloquent area gliomas who had no obvious limb dysfunction and 8 patients with non-eloquent area gliomas were admitted. All patients underwent task-based and resting state fMRI. Lateralization index (LI) of M1 activation was calculated. The regions-of-interest (ROIs) within motor network were selected. Functional connectivity (FC) between bilateral M1 and other ROIs were calculated. Occurrences of postoperative paralysis were recorded three months after operations. Receiver operating characteristic (ROC) analysis was performed to determine the optimum LI thresholds for prediction of paralysis. Compared with non-eloquent gliomas group, LI of M1 increased significantly (P=0.001) in eloquent gliomas group. Compared with patients without paralysis, LI of M1 in patients with paralysis increased significantly (P=0.011). ROC analysis showed the area under the curve (AUC) was 0.867. The best cut-off value for LI in the prediction of postoperative paralysis is 0.31 with sensitivity of 87.5%, specificity of 87.5%. Patients were then divided into groups of high and low risk of operation by LI ≥ 0.31. In patients with high-risk of operation, the tumors caused more damages to FC than in patients with low-risk of operation. Task-based and resting state fMRI can be combined together to better evaluate the risk of surgery and to clarify the mechanism of functional damage and its compensation.