<p>In rock engineering projects, it is essential to obtain accurate mechanical parameters of the rock mass to effectively predict stress–strain behaviour and failure characteristics, guiding construction decisions. The most common approach involves obtaining rock mechanical parameters via laboratory tests, followed by calculating rock mass parameters. However, existing methods for converting laboratory-tested rock parameters to rock mass parameters face several challenges, such as complex testing procedures, high costs, and long testing durations. To address these issues, this study investigates prediction methods for rock mass physical and mechanical parameters. Through extensive rock mechanical experiments, model tests, and numerical simulations, the relationships and influencing factors between compressive strength, elastic modulus, cohesion, internal friction angle, and wave velocity were identified. Based on these findings, a method using machine learning algorithms was developed to predict rock mass mechanical parameters from wave velocity, enabling real-time adjustment during construction. This approach eliminates several complex steps, such as lithology determination and water content testing, while maintaining accuracy, making it suitable for long-term use at construction sites.</p>