Tight conglomerate reservoirs exhibit abundant hydrocarbon resources，and multi-stage fracturing of horizontal wells is typically employed to construct three-dimensional（3D）complex fracture networks in low-permeability reservoirs for enhancing productivity. However，the strong heterogeneity of conglomerates leads to unique fracture propagation mechanisms，ambiguous influence patterns of dominant controlling factors，and significant challenges in fracturing design and fracture morphology prediction，thereby constraining efficient volumetric stimulation. A heterogeneous cubic conglomerate model containing spherical gravels was established using the continuum-discontinuum element method（CDEM）to simulate the 3D fracture propagation process during fracturing. Key parameters，including failure modes，gravel penetration rate，stimulated reservoir volume（SRV），and fractal dimension，were quantitatively characterized. The Pearson correlation coefficient method was applied to analyze correlations between seven influencing factors （e. g.，permeability，gravel strength，cementation surface strength，and horizontal stress difference）and fracture parameters. The weights of dominant controlling factors were determined using the coefficient of variation method，and a multi-parameter nonlinear evaluation model for SRVand fractal dimension was developed based on the least squares regression. The study demonstrates that stress characteristics are the primary controlling factors for SRVand the fractal dimension of fractures. Horizontal stress difference shows significant negative correlations with both SRV and fractal dimension（correlation coefficients：-0.563 9 and -0.611 7，respectively），with corresponding weights of 31.2% and 28.5%. Gravel content exhibits a strong positive correlation with gravel penetration rate. Gravel failure predominantly follows Mode-I failure， generating low-complexity fractures. However，increased gravel strength and reduced cementation surface strength induce gravel-bypassing behavior，promoting fracture bifurcation and enhancing fractal dimension. Additionally，fractal dimension is negatively correlated with injection displacement，and high injection displacement leads to the formation of a simplified fracture system dominated by the main fracture. Permeability exhibits higher sensitivity to fractal dimension（weight：11.5%）than to SRV（6.58%），reflecting constrained fracturing fluid filtration under low-permeability conditions，which inhibits fracture propagation scale，and fracture complexity is constrained by the flow capacity of reservoirs.