Computational modeling of fluid flow in heterogeneous and naturally fractured porous media is a fundamental tool for predicting subsurface flow behavior. However, achieving accurate and efficient numerical results remains a significant challenge. Classical cell-centered finite-volume schemes, such as the Two-Point Flux Approximation (TPFA), are robust, monotone, and computationally efficient for k-orthogonal meshes; however, they lose consistency on non-k-orthogonal meshes. Multi-Point Flux Approximation (MPFA) schemes overcome these limitations by providing consistent flux representations for general non-k-orthogonal meshes; however, they can produce solutions that violate the discrete maximum principle (DMP), leading to spurious oscillations and non-physical pressure fields. Fractures introduce further complexity, as they create strong permeability contrasts that can act as either flow channels or barriers. To address these challenges, this study develops a nonlinear flux-limiting strategy based on the Flux-Corrected Transport (FCT) approach to eliminate unphysical pressure oscillations and ensure DMP compliance in MPFA-type schemes. Our method applies a conservative algebraic flux-splitting formulation that constructs a monotone diffusive operator with M-matrix properties and a symmetric anti-diffusive counterpart that is applicable to any consistent MPFA scheme. The formulation is implemented in a hybrid modeling framework that couples non-orthodox MPFA with a Diamond Stencil (MPFA-D) for the rock matrix and the Continuous-Projection Embedded Discrete Fracture Model for Unstructured Meshes (CpEDFM-U) for explicit fracture representation. Benchmark tests from the literature confirm that our MPFA-D/CpEDFM-U/FCT formulation effectively eliminates spurious oscillations, preserves local mass conservation, and achieves accurate performance in modeling single-phase flow in heterogeneous and fractured porous media using general unstructured meshes.