Frosty Pod Rot (FPR) is caused by the fungus Moniliophtora roreri. This is a serious disease and can cause up to 80% of yield loss. Currently, the disease management control is tough: it consists in weekly removal pods and diseased pods must be buried. Until now, very few studies were conducted on the factors influencing pods infection, but it is known that climatic factors such as temperature, precipitation relative humidity and moisture play a role in this infection (Leandro-Muñoz 2011). This paper aims to study the effect of the pod production dynamic of the cocoa tree and inoculum sources on the FPR development in relation with the climatic factors. Three clones were studied: CATIE R4, CC137 and Pound 7 with differences of resistance to the disease. The results confirmed that CATIE R4 is the most resistant and Pound 7 the most sensitive. Bagging the diseased pods, which was equivalent to the elimination of the local inoculum on the plot, showed a significant difference. This result also confirmed previous results and supports the current disease control. But the local inoculum must be studied at a bigger scale and not only at the plot one. Our results suggested studying it at the plot scale and nearby plots scale. Our results also showed that the generation is crucial in the response to the disease: according to the generation, a clone can behave like another one, meaning be more resistant or sensitive. This point is of very importance and underlines the interest of studying cocoa phenology to find clones able to produce in climatic periods unfavourable to the fungus. Then, in order to understand how the climate influences changes of status of pods (from healthy to diseased), we constructed a binomial Generalized Linear Model (GLM) highlighting climatic key periods of the change of status (relatively to the date of labelling studied pods, the date and the duration of the climatic variable considered). We demonstrated that, according to the clone, key periods are