Host–parasite interactions between phytoplankton and fungi (chytrids) are key processes in aquatic ecosystems. However, individual-level heterogeneity in these interactions remains unexplored, although its importance in predicting the spread of diseases has been demonstrated in epidemiology. In this study, we experimentally tested whether individual-level heterogeneity could be a good indicator of phytoplankton–chytrid interactions, using a freshwater green alga Staurastrum sp., the diatoms Ulnaria sp. and Fragilaria crotonensis, and chytrid fungi. The number of attached fungi per host cell showed a non-random aggregated parasite distribution on Ulnaria sp. and F. crotonensis, but a random Poisson distribution on Staurastrum sp. To explore the potential mechanisms of these patterns, we developed a mathematical model describing sequential encounters between chytrid zoospores and host cells. As an equilibrium frequency distribution of the number of infections per host cell, we derived an exponentially-weighted Poisson distribution (EWPD) with a new weighting function. The statistical fits of the model explained the parasite distributions for Ulnaria sp. and F. crotonensis well, indicating that the clumped parasite distributions may result from the dependence of infection rate on the number of infections that already occurred on each host cell. Finally, additional incubation experiments using Ulnaria system supported this indication.