<p>Currently, most studies on predator–prey systems do not consider the dynamical behavior under the combined effects of higher-order interactions and diffusion. To better understand the complex ecological interactions between predators and prey, this paper proposes an improved Leslie-Gower reaction-diffusion predator–prey system with higher-order interactions. The system incorporates higher-order interactions, combining direct interactions such as fear effects, intraspecific competition, saturation effect, predator interference,and further introduces cross-diffusion. In the non-diffusive system, the conditions for the existence of a biologically feasible equilibrium point are given. The stability conditions for the equilibrium point are also determined. In the diffusive system, the conditions for Turing instability are presented, and the higher-order interaction coefficients are used as bifurcation parameters. Using a multiscale approach, the amplitude equations for two-dimensional Turing patterns at the Turing bifurcation threshold are derived. The correctness of the theoretical results is verified through numerical simulations. The results indicate that increasing the strength of higher-order interactions can lead to a transition from two-dimensional Turing patterns to spot mode, slow down the evolution speed of two-dimensional Turing patterns, and cause a transition of three-dimensional pattern modes to hollow tubular forms. Additionally, a sensitivity analysis of the system has been conducted, further confirming the significant impact of higher-order interaction mechanisms on the dynamic evolution of the system.</p>