<p>A probabilistic framework was developed to quantify landslide runout under nonstationary and cross-correlated spatial variability of soil strength. Nonstationary random fields with prescribed cross-correlation were constructed to represent cohesion and internal friction angle, and both parameters were assumed to increase linearly with depth. These random fields were incorporated into a material point formulation, and Monte Carlo simulations were conducted to evaluate the statistical characteristics of runout distance. The influences of key parameters, including the coefficient of variation, cross-correlation coefficient, vertical fluctuation scale, and strength gradient, were systematically examined. The results showed that circular slip failure dominated under both stationary and nonstationary conditions, and nonstationarity led to slightly higher mean values and dispersion of runout distance. Increased variability and negative cross-correlation between strength parameters amplified the uncertainty in runout. Larger vertical fluctuation scales and stronger depth-dependent strength gradients suppressed the development of deep slip surfaces and reduced runout distance. The proposed framework captured the coupled effects of nonstationarity and cross-correlation on large-deformation behavior and provided a quantitative basis for evaluating landslide impact zones.</p>