Integrating renewable energy sources (RES) into power systems introduces significant uncertainties that challenge static stability. Traditional deterministic methods fail to capture these stochastic effects, requiring probabilistic approaches. This paper shows a probabilistic power flow (PPF) analysis methodology in hybrid power systems incorporating RES. The impact of uncertainties from renewable generation and demand is evaluated using Monte Carlo Simulation (MCS) and Latin Hypercube Sampling (LHS). Machine learning (ML) techniques are used to fine-tune LHS sample size selection for enhanced computational efficiency, specifically the elbow detection method using the Euclidean norm versus sample size. The study applies these techniques to a 3-bus test system, the methodology shows the LHS’s ability to achieve comparable accuracy to MCS using only 650 samples, reducing computational time by over 90% (from 34.49 s to 3.17 s). Gaussian Mixture Models (GMMs) effectively capture multimodal uncertainties in generation and demand profiles. The results highlight LHS’s superior efficiency and low relative errors, making it ideal for large-scale systems, maintaining accuracy in capturing the probabilistic behavior of power flow variables such as voltage magnitudes and angles. This approach enhances risk assessment and supports sustainable energy system design by enabling rapid, accurate PPF analysis under high RES penetration.