Visual floral cues are used to enhance biodiversity and ecosystem services in forest landscapes by managing flowering resources in forest margins, rides, glades, and understory openings that support pollinators and beneficial insects. Yet field approaches for screening insect responses to colour cues are limited because observer presence and background stimuli can bias behavioural measurements. We developed a low-interference assay to quantify colour choice using artificial “flower dishes” baited with a honey-scented reward, and demonstrate it with the butterfly Aporia bieti (Lepidoptera: Pieridae) as a model nectar-foraging insect in a managed forest setting. The assay was deployed in an open meadow within Xiama Forest Farm (Gansu, China), representing a forest-associated clearing where insect-mediated plant reproduction contributes to ecosystem functioning. Colour stimuli were presented as discrete dishes; foraging responses (approach and visitation/feeding events) were video-recorded and scored offline to obtain consistent, disturbance-minimized readouts across observation periods. A. bieti approached and visited yellow stimuli more frequently than other colours, indicating strong attraction to yellow cues in this landscape. The dish design improved comparability across observation periods and reduced ambiguity in behavioural scoring, providing a practical framework for repeatable field measurements. For sustainable forest management, such measurements can inform selection and colour composition of flowering resources in forest edges and clearings, supporting pollination-related ecosystem services and forest biodiversity conservation. The framework can also support biodiversity surveillance by providing a low-cost behavioural readout comparable across sites and seasons. Because the assay is modular and inexpensive, it can be extended across taxa and combined with experience manipulations (pre-exposure or reward switching) to test experience-dependent modulation of colour choice under field conditions.