The article reviews the hardware implementation of memristive artificial neural networks (ANNs), focusing on the latest advancements and challenges. Memristors, a novel technology beyond CMOS, offer significant advantages in terms of energy efficiency and computational throughput due to their ability to store and compute simultaneously. The authors provide a comprehensive protocol for materials and methods involved in memristive neural networks, detailing the working principles of each block and design alternatives. They discuss the importance of high parallelization and near-memory computing to address the data communication bottleneck in conventional von Neumann machines. The article also covers the development of sophisticated ANNs, their applications in various fields, and the challenges of CMOS-based ANNs in terms of energy and area efficiency. The authors explore the use of memristive devices to emulate synapses, highlighting their potential to reduce power consumption and improve integration density. They describe the structure of memristor-based ANNs, including image capture hardware, input vector conformation, input driving circuits, vector-matrix multiplication (VMM) core, and sensing electronics. The article provides a detailed explanation of the VMM core, emphasizing the parallel processing capabilities of memristor crossbars and the strategies for representing signed weights. Finally, it discusses the sensing electronics used to measure the output currents from the memristor crossbar.The article reviews the hardware implementation of memristive artificial neural networks (ANNs), focusing on the latest advancements and challenges. Memristors, a novel technology beyond CMOS, offer significant advantages in terms of energy efficiency and computational throughput due to their ability to store and compute simultaneously. The authors provide a comprehensive protocol for materials and methods involved in memristive neural networks, detailing the working principles of each block and design alternatives. They discuss the importance of high parallelization and near-memory computing to address the data communication bottleneck in conventional von Neumann machines. The article also covers the development of sophisticated ANNs, their applications in various fields, and the challenges of CMOS-based ANNs in terms of energy and area efficiency. The authors explore the use of memristive devices to emulate synapses, highlighting their potential to reduce power consumption and improve integration density. They describe the structure of memristor-based ANNs, including image capture hardware, input vector conformation, input driving circuits, vector-matrix multiplication (VMM) core, and sensing electronics. The article provides a detailed explanation of the VMM core, emphasizing the parallel processing capabilities of memristor crossbars and the strategies for representing signed weights. Finally, it discusses the sensing electronics used to measure the output currents from the memristor crossbar.