Noise pollution is a significant challenge for achieving sustainable environments, causing adverse health impacts and economic losses. Urban planners seekways to design quieter neighbourhoods, but conventional methods cannot guarantee the perceived quality of sound environments. Soundscape has emerged as a promising framework for defining and predicting the acoustic environment as perceived by the people, but it still lacks efficient implementation in practice. In this work, we have explored an approach towards the soundscape design of a small urban park by reducing noise via sound-manipulating acoustic metamaterial (AMM) structures. We collected real-time traffic noise data in a small urban park in Central London and investigated the influence of AMM structures on psychoacoustic metrics such as loudness, sharpness, and roughness, which are key for soundscape analysis. Our results revealed that the presence of AMM structures as the size of a bench trained on a single frequency (200Hz) influences the sharpness (S), roughness (R), and loudness (N5) values by 10.6%, 11% and 18.2%, respectively. In turn, this can potentially increase the pleasantness and eventfulness of soundscape, a stepping stone in building efficient noise-free interventions to enrich our lives and foster a healthier environment.