Commodity plastics such as high density polyethylene (HDPE) have become integral to society. However, the potentially long-lasting ecological impacts of these plastics have spurred researchers to search for more sustainable solutions. One such solution is to develop a method for designing plastics with tunable and improved properties, thus decreasing the amount of material needed for various applications. In this work, we report a machine learning approach that maps the relationship between polymer molecular weight distributions (MWDs) and the physical properties (tensile and rheological) of HDPE. Using this approach, we design and generate HDPE materials with user-specified properties and valorize degraded postconsumer polyethylene waste. Implementation and development of this approach will facilitate the design of next-generation commodity materials and enable more efficient polymer recycling, thereby lowering the overall impact of HDPE on the environment.