Current research has not fully explored how streetscape elements in different street spaces affect long-distance, high-speed recreational cycling. As a result, the applicability of existing findings across different street environments is limited, hindering their practical value in urban street design. To address this issue, this study focuses on the core urban area of Hangzhou, China. Streets are functionally categorized based on Point of Interest (POI) data, and Strava crowdsourced data are used to obtain cycling activity trajectories. Streetscape indicators that may influence recreational cycling paths are calculated using streetscape imagery and machine learning techniques. Ordinary Least Squares (OLS) and Geographically Weighted Regression (GWR) models are employed to identify significant streetscape indicators that impact cycling, and spatial heterogeneity in cycling route preferences is analyzed at different scales across street types. The results indicate that: (1) Recreational cyclists prefer mixed-use and scenic streets, particularly those along river corridors or within and around large green spaces. (2) Among the streetscape indicators, safety isolation degree has the most positive impact, followed by facility diversity, motorization level, and green view index, all of which positively influence cycling traffic to varying extents. Interface transparency and disorder negatively affect cycling, with transparency having the strongest influence. (3) In mixed-use, commercial, and life service streets, the effect of streetscape indicators on cycling route preferences varies by spatial location, exhibiting both promoting and inhibiting effects. In scenic streets, streetscape indicators show stronger positive or negative impacts, while in industrial streets, their influence is weaker. Based on these findings, the study proposes strategies for creating cycling-friendly environments tailored to different street types. The results validate and extend existing theories on the interaction between street environments and cycling behavior, offering valuable insights for diagnosing problems and implementing effective interventions to promote cycling-friendly environments, thus contributing to urban health, equity, and sustainable development.