In an era of growing demand for high-performance systems, it is essential to develop tools that keep pace with technological advancements. Heterogeneous programming, combined with the C++ language, represents a crucial step towards the efficient use of computational resources in technical and scientific fields such as parallel computing, computational simulations, and the development of artificial intelligence models. This paper presents the fundamentals of heterogeneous systems and heterogeneous programming, along with a detailed review of the SYCL standard and its implementation through the oneAPI DPC++ compiler. The process of setting up a local SYCL environment is explained, as well as the possibilities of using ready-made cloud environments. Basic examples are provided to illustrate fundamental SYCL concepts and classes, as well as methods for measuring the execution time of SYCL kernels. Matrix convolution demonstrates how to link C++ multithreading and the execution of SYCL programs on multiple accelerators. The introduction of the ND-Ranges paradigm explains the difference between global and local memory on the device, as well as synchronization methods for work items within the same subgroup. Finally, the oneAPI DPC++ library is described, which facilitates writing heterogeneous programs through a higher level of abstraction and integrated parallel mechanisms. SYCL's code portability and the ability to provide a unified approach to various accelerators significantly simplify the development of heterogeneous applications, but they require a careful understanding of hardware architecture to achieve optimal results. While challenges, such as the complexity of setting up environments and adapting to specific hardware, remain, SYCL represents a powerful tool for developing future heterogeneous systems. With the continuous development of the SYCL standard and the compilers that implement it, its presence will undoubtedly grow, primarily due to reduced complexity compared to other similar standards and its support for accelerators from various manufacturers.