This article presents a novel two-terminal ferroelectric fin diode (FFD) designed to combine the functionalities of both digital and analog memory devices. The FFD consists of a ferroelectric capacitor and a fin-like semiconductor channel, sharing both top and bottom electrodes. This design offers robustness and universality, as it can operate with different ferroelectric materials, including organic and inorganic compounds. The FFD demonstrates superior performance compared to existing non-volatile memories, with an endurance of up to 10^10 cycles, an ON/OFF ratio of -10^4, a feature size of 30 nm, an operating energy of -20 fJ, and an operation speed of 100 ns. The simple two-terminal structure and high self-rectification ratio of -10^4 make it suitable for designing passive crossbar arrays, which are crucial for in-memory computing applications. The device's uniformity and analog storage capabilities are highlighted, with a device-to-device variation ratio of -0.023 for positive coercive voltage and -0.019 for negative coercive voltage. The FFD's potential for in-memory computing is demonstrated through a pattern classification task using a 1.6 k unit passive crossbar array, achieving high recognition accuracy. This work opens new avenues for efficient memories and emerging computing architectures in big data and artificial intelligence applications.This article presents a novel two-terminal ferroelectric fin diode (FFD) designed to combine the functionalities of both digital and analog memory devices. The FFD consists of a ferroelectric capacitor and a fin-like semiconductor channel, sharing both top and bottom electrodes. This design offers robustness and universality, as it can operate with different ferroelectric materials, including organic and inorganic compounds. The FFD demonstrates superior performance compared to existing non-volatile memories, with an endurance of up to 10^10 cycles, an ON/OFF ratio of -10^4, a feature size of 30 nm, an operating energy of -20 fJ, and an operation speed of 100 ns. The simple two-terminal structure and high self-rectification ratio of -10^4 make it suitable for designing passive crossbar arrays, which are crucial for in-memory computing applications. The device's uniformity and analog storage capabilities are highlighted, with a device-to-device variation ratio of -0.023 for positive coercive voltage and -0.019 for negative coercive voltage. The FFD's potential for in-memory computing is demonstrated through a pattern classification task using a 1.6 k unit passive crossbar array, achieving high recognition accuracy. This work opens new avenues for efficient memories and emerging computing architectures in big data and artificial intelligence applications.