To mitigate the trade-off between food security and ecological security is a big challenge that the human-beings are facing nowadays. To this aim, land use optimisation is often conducted based on the superiority of the land to provide food or ecosystem services. However, previous studies largely ignore cropland use intensity in calculating food production and normally only consider the quantity and quality of the cropland. This is inflexible in accessing food security for land use optimisation as cropland use intensity and other constraints change. This study considers cropland use intensity for the prediction of food production. The Yangtze River Economic Belt was taken as the study area and a binary linear programming model is applied to optimise land use pattern in 2020. A Geographically Weighted Regression model is used to calculate food production based on the cropland use intensity concerning fertiliser, pesticide and agricultural film at a city level. The ecological security is evaluated according to the ecosystem service values modified by ecological sensitivity. Accordingly, the land use pattern is optimised to maximise ecological security and ensure food production remaining no less than that before optimisation. Results indicate an increase of ecological security at 2.10% without the cost of food security after land use optimisation. The area of cropland generally increases in the lower reach, while an increase of ecological land in the upper reach of the Yangtze River Economic Belt. This study highlights the incorporation of cropland use intensity in accessing food production for land use optimisation to coordinate food security and ecological security.