Objective: Mild cognitive impairment (MCI) is often a transitional state between normal aging and Alzheimer’s disease (AD). When observed longitudinally, some MCI patients convert to AD, while a considerable portion either remains MCI or revert to a normal functioning state. This divergence has provided some enlightenment on a potential biomarker to be represented in the resting state brain activities of MCI patients with different post-hoc labels. Recent studies have shown impaired executive functions, other than typically explicated memory impairment with AD/MCI patients. This observation raises the question that whether or not the executive control network (ECN) was impaired, which pivotally supports the central executive functions. Given the fact that effective connectivity is a sufficient index in detecting resting brain abnormalities in AD/MCI, the current study specifically asks a question whether the effective connectivity patterns are differentiated in MCI patients with different post-hoc labels.Methods: We divided the MCI subjects into three groups depending on their progressive state obtained longitudinally: 1) 15 MCI-R subjects: MCI reverted to the normal functioning state and stabilized to the normal state in 24 months; 2) 35 MCI-S subjects: MCI patients maintained this disease in a stable state for 24 months; 3) 22 MCI-P subjects: MCI progressed to AD and stabilized to AD in 24 months, and 4) 39 age-matched normal control subjects (NC). We conducted a Granger causality analysis after identifying the core nodes of ECN in all of the subjects using Independent Component Analysis. Our findings revealed that different MCI groups presented different effective connectivity patterns within the ECN compared to the NC group. Specifically, (1) dorsolateral prefrontal cortex (dLPFC) and medial prefrontal cortex (mPFC) were the core nodes in the ECN network that exhibited different connecting patterns; (2) an effective connection circuit “R.dLPFC→ right caudate→ left thalamus→R.dLPFC” in the ECN showed different levels of damage; and (3) there were four pathways between the R.dLPFC and L.LP, and these four pathways were also different. Results: Our results would help to understand the potential central mechanism of MCI patients. The differentiated effective connectivity of ECN may serve as a potential biomarker for early detection of AD, which may also provide a reference for clinical researchers to manipulate active but distinctive interventions for MCI patients who have different risks.