Abstract: The accurate and rational classification of rock mass blastability is a critical prerequisite for optimizing blasting parameter design and reducing excavation costs in mining engineering. However, the evaluation of rock mass blastability is inherently characterized by significant uncertainty and randomness due to the complex geological conditions of rock masses and nonlinear interactions among influencing factors. In addition, conventional weighting approaches based on a single method often suffer from either excessive subjectivity or substantial objective bias, which may reduce the reliability of evaluation results. To address these limitations, this study proposes a rock mass blastability evaluation method based on the integrated weighting and multidimensional connection cloud model. First, an evaluation index system for rock mass blastability was established by selecting four representative indicators: uniaxial compressive strength of rock (σ), rock density (γ), rock mass integrity coefficient (η), and specific explosive consumption (q). To improve the rationality of indicator weighting, the improved analytic hierarchy process and entropy weight method were employed to calculate subjective and objective weights, respectively. Considering the limitations of relying solely on either subjective expert judgment or objective data dispersion, a distance function was introduced to integrate the two weighting results to obtain more balanced and reliable comprehensive weights for evaluation indicators. On this basis, the traditional multidimensional cloud model was further improved by incorporating set pair analysis theory to establish a multidimensional connection cloud model. This improvement effectively addresses the ambiguity of grade boundaries in conventional rock blastability classification and enhances the model’s ability to characterize both fuzziness and randomness in evaluation processes. Using the improved model, 29 groups of measured engineering data were analyzed for indicator attribute assessment. Combined with a multidimensional forward cloud generator, the comprehensive membership degree of each rock sample to different blastability grades was calculated. Subsequently, the blastability grade of each sample was determined according to the maximum membership principle. To further investigate the intrinsic relationships between rock mass blastability and evaluation indicators, Kendall correlation analysis was conducted. The results indicate that the blastability classification outcomes obtained using the proposed method are highly consistent with those generated by a back propagation neural network model, with an overall classification accuracy of approximately 90%, demonstrating the reliability and effectiveness of the proposed approach. The Kendall correlation analysis reveals that the uniaxial compressive strength exhibits the strongest correlation with rock mass blastability, followed by the rock mass integrity coefficient and rock density, while specific explosive consumption shows the weakest correlation. The correlation coefficients among different indicators are mainly concentrated within the range of 0.03–0.26, indicating a relatively weak interdependence. This suggests that the selected indicators can independently characterize rock mass properties from different dimensions to validate the rationality of the constructed evaluation index system. The findings of this study provide theoretical support for the design and optimization of blasting parameters in rock engineering projects. More importantly, the proposed method offers a new perspective and methodological framework for rock mass blastability evaluation under conditions of uncertainty, which may contribute to safer, more efficient, and cost-effective mining operations.