A phenology model of the walnut husk fly, Rhagoletis completa Cresson, was developed to more accurately predict the timing of the flight period and optimize management decisions. A data set of 153 orchard years in which adults were trapped throughout the season was used for the development and validation of this model. Data from California Irrigation Management Information System (CIMIS) weather stations were used to match orchard-year datasets with historical climatic data on degree-day (DD) accumulation, winter chill, and winter rainfall. A cumulative Weibull distribution was used to model the relationship between cumulative trap catch and DD accumulation for R. completa in California. The model was used to predict thermal requirements for the start (5% cumulative trap catch) and mid-point (50% cumulative trap catch) of the flight period, which were 1,670 and 2,179 DDs, respectively. The prediction for 50% cumulative trap catch of R. completa in California was much higher than the thermal requirement estimated in Oregon previously (1,751 DDs). Linear mixed effects models were used to evaluate other environmental and orchard-specific factors which could explain the large variation between predicted and observed thermal requirements for both the start and mid-point of the flight period. Latitude, walnut cultivar leaf-out time, orchard age and year, as a continuous variable, all contributed significantly to explain deviations from the predictions of the DD model for individual orchard years. Such factors can be used both to adjust predicted thermal requirements for these two specific and informative stages of the flight period, and to provide a basis for ecological and evolutionary hypotheses.