Rice spikelets become sterile when exposed to high temperatures during flowering. Recent experimental studies have shown that transpirational cooling and flowering time (within the day) are important factors determining sterility. Extrapolation of the results from these experimental studies to other environments requires an integrated approach through incorporating such effects on spikelet fertility in a crop growth model. Here we review existing fertility functions (fertility as a function of panicle temperature) in different studies, and propose a simple new model that incorporates insights on diurnal patterns of temperature and relative humidity, recent studies on the relation between panicle temperature and sterility, and studies on how flowering time depends on environmental conditions. We illustrate the model for a typical arid and typical humid climate. There was a large difference between the existing fertility functions, and simulated fertility was very sensitive to differences in fertility functions, causing differences in simulated fertility of up to 59%. Ignoring transpirational cooling led to overestimation of sterility of 14–73%. Shifting flowering times from 12:00 to 9:00 led to a 7–35% reduction in sterility. Within day flowering duration had only a marginal impact (max 2%). We conclude that any climate change impact simulation is highly dependent on the choice of the fertility function, and identification of causes of differences between reported fertility functions deserves further investigation.