The discrimination of liquors plays a vital role in our daily life as a large amount of cheap alcoholic beverages and low-quality counterfeits greatly damage the health of consumers. Array-based optical sensors have been developed as a fast and effective tool to discriminate multi-analytes or complex mixtures with similar structures, such as acids, esters, aldehydes and ketones. Over the past decades, optical sensor arrays have demonstrated remarkable capabilities in fast and accurate detection of diverse types, qualities, and authentic liquors. The identification strategy mainly focuses on the specific/non-specific interactions of flavor compounds in liquors with sensor elements. The cross-reactive optical signal readouts from the combination of sensor elements with liquors and multiple machine learning algorithms are essential to ensure the discriminatory capability of sensing system. In this review, we discuss the advantages and recent advances in pattern recognition of liquors in terms of the design of various types of sensing materials and the construction strategies of sensor arrays.