One of the fundamental goals of ecology is to examine how dispersalaffects the distribution and dynamics of insects across natural landscapes. These landscapes are frequently divided into patches of habitat embedded in a matrix of several non-habitat regions, and dispersal behavior could vary within each landscape element as well as the edges between elements. Reaction–diffusion models are a common way of modeling dispersal and species interactions in such landscapes, but to apply these models we also need methods of estimating the diffusion rate and any edge behavior parameters. In this paper, we present a method of estimating the diffusion rate using the mean occupancy time for a circular region. We also use mean occupancy time to estimate a parameter (the crossing probability) that governs one type of edge behavior often used in these models, a biased random walk. These new methods have some advantages over other methods of estimating these parameters, including reduced computational cost and ease of use in the field. They also provide a method of estimating the diffusion rate for a particular location in space, compared to existing methods that represent averages over large areas. We further examine the statistical properties of the new method through simulation, and discuss how mean occupancy time could be estimated in field experiments.