In many urban areas, use of herbicides is either unwanted or prohibited and replaced with flame weeding. The influence of dose (kg propane ha-1) and treatment interval of flame weed control was studied on Lolium perenne. Lolium perenne is a perennial grass that is very difficult to control with non-chemical weed control methods, because of its extensive regrowth. Treatments of eight different doses and five treatment intervals were applied during two seasons from May to October. The response was measured as plant dry weight, 14 days after last treatment. All weeds were killed with doses above 80 kg propane ha-1 when treatments were carried out every other week (10 treatments per growing season and a total dose of 800 kg ha-1), and no regrowth was seen the following 2 weeks. Six treatments a year and a total dose of 631–674 kg ha-1, depending on year, controlled L. perenne effectively (90% reduction in dry weight). Split applications generally increased the effect of the treatments, especially when the number of treatments was increased from four to six. The results are in accordance with the assumption that repeated flame treatments are necessary to kill larger plants and heat tolerant weeds, such as grasses and perennial weeds that will regrow after a single treatment. Knowledge of the relation between dose and treatment intervals may be used to improve flame weeding strategies on hard surfaces.