The number of interconnected equipment is predicted to surge to 75 billion by 2025, and the involved global data will swell to 175 zettabytes. [3,4] The massive data threaten the prevailing von Neumann architecture, whose processing and memory units are physically separated, resulting in low efficiency when handling artificial intelligence (AI) tasks. [5] Process technologies also restrict the development of processors and memorizers along the path of Moore's law, seriously restricting hardware computational power and storage capacity. Although complementary metal-oxide-semiconductor (CMOS) has played a crucial role in modern civilization, it is difficult to satisfy the demand of Big Data and AI era. [6] Therefore, it is inevitable to develop a new chip architecture and device technology. The inspiration for new technology comes from the neural network that is constituted of 10 11 neurons and 10 15 synapses. [7] On this basis, the human brain can not only learn and memorize information autonomously, but also recognize, reason, and process multiple tasks simultaneously, which features low energy consumption, high stability, and parallel processing. Therefore, the computing and memory devices mimicking brain have drawn extensive attention in recent years. [8,9] Among these emerging devices, memristors are expected to realize neuromorphic processors characterizing in situ in-memory computations using networks-on-chip in an event-driven manner. [10] By virtue of the history of the amount of flowing charges, memristors can be applied to artificial neural network (ANN) and data memory. [11][12][13] Flexible memristors can satisfy the requirements of portable, wearable, and implantable applications. [14] The controllability of memristors can ensure data-processing accuracy and stability. Specifically, the memristance modulation can manifest as the mollification of spatial (device-to-device) and temporal (cycle-to-cycle) variability or the realization of interconversions between different resistive switching behaviors. [15,16] The classical memristor consists of top and bottom electrodes and an intermediate functional layer (Figure 1a), which is an easy integration with CMOS and highly extensible two-terminal device. [17][18][19][20] Functional layer is the heart of memristor, and the related materials determine its performance. Since the first physical model of memristor was reported using TiO 2 as a functional layer, many materials, including carbon dots (CDs), have exhibited good memristive performances. [21][22][23] As a kind of carbon material, CDs generally possess remarkable characteristics, including physicochemical stability, The explosive growth of digital communication promotes advanced memory and computing devices in Big Data and artificial intelligence era. In particular, memristors hold great promise for in-memory computing and artificial synapses, expected to break through restrictions on hardware computational power and storage capacity caused by the von Neumann bottleneck and declining Moore's Law. The m...