<p>Dynamic accuracy is critical for industrial applications of hybrid robots. Although existing research has extensively investigated joint clearance effects, systematic methods for quantitatively evaluating individual clearance joint importance remain lacking, particularly where multi-joint interactions complicate analysis. Moreover, as modeling extends to component flexibility and wear effects, computational burden increases dramatically, constraining parametric optimization and real-time control. Therefore, developing reduced-order methods balancing fidelity with efficiency has become urgent. This paper analyzes dynamic responses of a 5-DOF hybrid robot considering clearance and flexibility effects, proposing a Morris screening-based model order reduction method. A rigid-flexible coupling model is established using Cartesian coordinates for rigid components, ANCF gradient-deficient beam elements for flexible components, and embedded clearance models. Based on industrial pick-and-place trajectories, influences of clearance magnitudes and motion durations are systematically analyzed. Results show clearance and flexibility effects significantly impact performance, with distinct forward-return path differences and substantial joint contribution variations. Morris sensitivity analysis identifies critical clearance joints, revealing dynamics are dominated by limited high-sensitivity joints. Model reduction retains critical joints while simplifying low-sensitivity joints as ideal constraints. Validation demonstrates 40% computational reduction while maintaining displacement-velocity accuracy and capturing critical acceleration regions, effectively balancing fidelity with efficiency for complex rigid-flexible systems.</p>