Human activities put stress on our oceans and with a growing global population, the impact is increasing. Stressors rarely act in isolation, with the majority of marine areas being impacted by multiple, concurrent stressors. Marine spatial cumulative impact assessments attempt to estimate the collective impact of multiple stressors on marine environments. However, this is difficult given how stressors interact with one another, and the variable response of ecosystems. As a result, assumptions and generalisations are required when attempting to model cumulative impacts. One fundamental assumption of the most commonly applied, semi-quantitative cumulative impact assessment method is that a change in modelled cumulative impact is correlated with a change in ecosystem condition. However, this assumption has rarely been validated with empirical data. We tested this assumption using a case study of seagrass in a large, inverse estuary in South Australia (Spencer Gulf). We compared three different seagrass condition indices, based on survey data collected in the field, to scores from a spatial cumulative impact model for the study area. One condition index showed no relationship with cumulative impact, whilst the other two indices had very small, negative relationships with cumulative impact. These results suggest that one of the most commonly used methods for assessing cumulative impacts on marine systems is not robust enough to accurately reflect the effect of multiple stressors on seagrasses; possibly due to the number and generality of assumptions involved in the approach. Future methods should acknowledge the complex relationships between stressors, and the impact these relationships can have on ecosystems. This outcome highlights the need for greater evaluation of cumulative impact assessment outputs and the need for data-driven approaches. Our results are a caution for marine scientists and resource managers who may rely on spatial cumulative impact assessment outputs for informing policy and decision-making.