In data mining, missingness is a serious challenge when dealing with real-world data sets. Although several approaches have been proposed to tackle missing values in machine learning, most studies focus on the classification task rather than the regression task. To the best of our knowledge, no study has been conducted to investigate the Symbolic Regression(SR)on incomplete real-world data sets. This work aims to conduct a research on using Genetic Programming (GP) for SR with data sets suffering from the presences of incompleteness. This implies developing new methods to handle missing values for SR. Moreover, some data challenges such as high-dimensionality and scalability will be addressed. To achieve the objectives of this thesis, several methods will be developed by utilizing machine learning techniques such as feature manipulation, instance selection, and transfer learning. Such methods will be evaluated on synthetic and real-world incomplete data sets.