Passive acoustic monitoring is increasingly used to assess biodiversity across broad spatial and temporal scales, particularly in systems where traditional surveys are logistically challenging. Acoustic indices, such as the Acoustic Complexity Index (ACI), Bioacoustic Index (BI), and Normalized Difference Soundscape Index (NDSI), are widely applied as proxies for biological diversity, community activity, or ecosystem condition. However, these indices are typically calculated without species-level information, and the extent to which they are influenced by dominant or invasive species remains insufficiently understood. In this study, we investigated how the invasive cane toad ( Rhinella marina ) influences acoustic indices across agricultural landscapes on Ishigaki Island, Okinawa, Japan. Continuous environmental recordings were collected from 16 sites encompassing three land-use types (pasture, paddy, and sugarcane fields), resulting in 29 site–season samples (13 in summer and 16 in winter). Recordings were obtained at 3-minute intervals over multiple weeks at each site. To focus on peak anuran activity, analyses were restricted to nocturnal periods (19:00–04:00). Cane toad vocalizations were manually identified with high accuracy using Kaleidoscope Pro, and their occurrence and abundance were quantified for each recording. In parallel, acoustic indices were calculated for all recordings independently of species identification, allowing direct comparison between species-specific activity and community-level acoustic metrics. Cane toad activity showed a strong association with land use. Daily detection counts were significantly higher in paddy fields than in pasture and sugarcane sites, indicating that paddy habitats function as core activity areas for this species. Seasonal differences were relatively weak at the landscape scale, with paddy fields maintaining high activity across both summer and winter. In contrast, pronounced heterogeneity was observed among sites. Even within paddy habitats, some sites exhibited intense activity in summer but minimal activity in winter, whereas others showed the opposite pattern. This indicates that seasonal hotspots of cane toad activity shift among sites, suggesting that fine-scale environmental conditions, such as local hydrology, microhabitat structure, or breeding-site availability, play a key role in determining site-level activity patterns beyond broad land-use categories. At the recording level, the presence of cane toads had consistent and significant effects on multiple acoustic indices. Recordings in which cane toads were detected exhibited higher values of ACI, BI, and NDSI compared to recordings without detections. These indices are commonly interpreted as indicators of increased biological diversity or acoustic complexity. However, our results demonstrate that such increases can arise from the activity of a single dominant species, rather than from changes in community-level biodiversity. This finding highlights a critical limitation in the interpretation of acoustic indices, particularly in ecosystems where invasive or highly vocal species are present. Our study provides two key insights. First, invasive species can influence soundscape structure not only through ecological interactions but also by altering how biodiversity is quantified using acoustic metrics. Second, the spatial and seasonal dynamics of invasive species activity can be highly heterogeneous even within a single land-use type, emphasizing the importance of site-level processes in understanding invasion patterns. Overall, these results demonstrate that acoustic indices should be interpreted with caution when used as proxies for biodiversity, and that integrating species-specific information is essential for robust ecological inference from soundscape data in invaded ecosystems.