One of the main problems in the consumption of edible flowers is the fact that those possess high perishability post-harvest, and it difficult its commercialization and consumption. For this reason, it is important to apply conservation methods to avoid losses, and one of these methods is the dehydration. This work aims to use mathematical regression models to describe the behavior of the flower Viola × wittrockiana during its drying period. Ten non-linear regression models were adjusted to the data set, which was obtained by experimentation in a 30°C temperature. The study concluded that the model that was the foremost adjusted was the Midilli et al., and that, with the assistance of this model, it is possible to predict the weight loss of the flower during its dehydration.Um dos principais problemas no consumo de flores comestíveis é o fato de que elas possuem alta perecibilidade pós-colheita, o que dificulta a sua comercialização e consumo. Por isso, é importante aplicar métodos de conservação para evitar perdas, e um desses métodos é a desidratação, comumente descrita pela cinética de secagem por meio de uma curva de secagem. O objetivo desse trabalho foi utilizar modelos matemáticos de regressão de modo a descrever o comportamento da flor Viola × wittrockiana durante a sua secagem. Dez modelos de regressão não-linear foram ajustados ao conjunto de dados de secagem obtido através de experimentação em uma temperatura de 30°C. Concluiu-se que o modelo que melhor se ajustou foi o de Midilli et al., e que, com o auxílio dele, pode-se prever a perda de massa da flor durante sua desidratação. Palavras-chave: regressão não-linear; secagem de flores; flores comestíveis; Viola × wittrockiana.   Non-linear models applied to the drying kinetics of pansy edible flower   ABSTRACT: One of the main problems in the consumption of edible flowers is the fact that those possess high perishability post-harvest, and it difficult its commercialization and consumption. Therefore, it is important to apply conservation methods to avoid losses, and one of these methods is dehydration, commonly described by drying kinetics through a drying curve. This work aims to use mathematical regression models to describe the behavior of the flower Viola × wittrockiana during its drying period. Ten non-linear regression models were adjusted to the data set, which was obtained by experimentation in a 30°C temperature. The study concluded that the model that was the foremost adjusted was the Midilli et al., and that, with the assistance of this model, it is possible to predict the weight loss of the flower during its dehydration. Keywords: non-linear regression; flower drying; edible flowers; Viola × wittrockiana.