Herbicide dose optimisation, i.e. maximising weed control and crop yield with herbicide dose, is an important part of integrated weed management strategies. However, the adoption of optimised dose technology and variable rate application has been limited because of the relatively long period between herbicide treatment and the time when efficacy can be visually assessed. Herbicide dose optimisation could therefore benefit from simple methods that allow early prediction of plant mortality. Early prediction would allow better management decisions, e.g. timely retreatment in case of uncontrolled weeds. The focus of this study was the relationship between leaf photosynthesis soon after herbicide treatment and subsequent plant mortality, with the aim of determining whether the former could predict the latter. Data from 28 glasshouse experiments were analysed. In these experiments, herbicides from five modes of action groups were tested on five plant species. Leaf photosynthesis was measured with two mobile meters up to 1 week after herbicide treatment. Leaf photosynthesis was affected by plant species, leaf number, herbicide species, dose and time. Large changes in leaf photosynthesis were observed with photosynthesis-inhibiting herbicides, intermediate changes were noted with glyphosate, glufosinate-ammonium and sulcotrione, and no changes were detected with MCPA. Threshold values associated with plant mortality were then determined. These values can be used to assess the risk of uncontrolled weeds treated with variable herbicide doses.