With the global population aging, early detection of mild cognitive impairment (MCI) as an initial stage of Alzheimer’s disease (AD) and other neurodegenerative conditions is crucial for cognitive health research. This review systematically summarizes various technical strategies for early MCI detection and their limitations, including cognitive testing, neuroimaging, eye tracking analysis, gait testing, and biomarker detection. Through comparative analysis, we highlight the potential of facial expression recognition (FER) technology as a nonintrusive and economically efficient screening tool. Recent advancements in convolutional neural networks (CNNs) and transformer architectures have significantly improved FER technology, which not only elevates classification performance but also facilitates dynamic parameter optimization under real environmental conditions. However, FER in MCI detection faces challenges, including short duration of micro-expressions, low expression intensity, and large inter-individual differences. Future directions include integrating multimodal approaches and improving dataset diversity to enhance model robustness. This review provides a comprehensive overview of FER technology’s role in MCI detection, summarizing existing research and identifying future opportunities.