This study uses the data-driven machine-learning technique called ridge regression to address an innovative method of designing hourglass lattice-structured metamaterials. Metamaterials are engineered materials with properties derived from their intricate structural arrangements, holding promise for various applications. The hourglass lattice arrangement is exciting because of its unique mechanical characteristics and possible advantages in stiffness m odulation. D esigning s uch c omplex s tructures o ften i nvolves m anual i terations a nd simulations, which can be time-consuming and limited in exploring the vast design space. In this paper, we suggest an innovative approach that uses data-driven machine learning to speed up and improve the design process. By training models on a dataset of metamaterial behaviors, we enable the prediction of optimal hourglass lattice configurations f or d esired m echanical p roperties. T his p rediction u ses a m achine l earning a lgorithm t o analyze the data obtained from existing design simulations. This predictive capacity empowers researchers and engineers to explore an extensive design space efficiently, thus uncovering optimal configurations that might remain undiscovered using traditional methods like manual adjustments and iteration or physical prototyping and testing, which are time-consuming and labor-intensive. Engineers iteratively refine d esigns b ased o n s imulation o r test results, limiting design space exploration and potentially missing optimal configurations. T he c ore innovation lies in the ability of these models to predict the mechanical properties and behaviors of hourglass lattice metamaterials based on their structural characteristics, such as radius of curvature and thickness. This methodology can potentially revolutionize metamaterial design by efficiently us ing da ta-driven ma chine-learning models.