In a farm, a fire that starts in any structure can spread to all other structures of that same farm, to barns, granaries, silos, etc.. Intuitively, we then expect that a farm with more structures will be more at risk of fire propagation than a smaller one. From an actuarial perspective, as the total premium for farm insurance is the sum of premiums of each structure of that farm, it is therefore necessary to propose a way to compute each premium by considering the risk of fire propagation. Based on the distances between structures on the same farm, we propose a new pricing approach that considers fire propagation. The proposed model makes it possible to analytically compute the probability of fire propagation as a function of the fire origin. This can in turn be used to price all individual structures of any given farm. A practical application of the model based on insurance data and satellite images is given.