Accurate knowledge of the internal core distribution of multicore fibres (MCFs) is essential, given their widespread application, including in fibre splicing, bundle fan-in/fan-out, mode coupling, writing gratings, and fibre drawing. However, the extensive use of MCFs is restricted by the limited methods available to precisely measure the fibre core distribution, as the measurement accuracy determines the performance of the product. In this study, a sideview and nondestructive scheme based on Bessel beam illumination was proposed for measuring the internal core distribution of a seven-core fibre. Bessel beams offer a large focal length in a scattering medium, and exhibit a unique pattern when propagating in an off-axis medium with a spatially varying refractive index. The results revealed that a long focal length and unique pattern influence the image contrast in the case of Bessel beams, which differs from a typical Gaussian beam. Further, high-precision measurement of a seven-core fibre core distribution based on a Bessel beam was demonstrated using a digital correlation method. A deep learning approach was used to improve the measurement precision to 0.2° with an accuracy of 96.8%. The proposed sideview Bessel-beam-based method has the potential to handle more complex MCFs and photonic crystal fibres.