Performance enhancement is the main wish in any industry. In the drilling process, the challenge lies in finding the right conditions to reach a desired depth faster, while balancing the operational complexities with the associated risks. In this regard, drilling operations generate enormous quantities of data and metadata with the main goal of providing detailed visualization of operations accessible remotely and in real time. This aligns with the existent big-data time, where data mining techniques appear as means to drive proficiencies in data processing to generate new and valuable information. From this perspective, the ultimate goal of this thesis is to assess the application of data mining software to transform commonly acquired drilling data into actionable data with possible impact in well planning and during later operations. In order to achieve the prime goal of the thesis the Rate of Penetration (ROP) was selected to be the focus of the study. The ROP, known as one of the contributors in time estimation for operations, is the variable of interest for the analysis and prediction. This work applies data mining techniques to examine pre-existing data sets of previously drilled wells looking for meaningful information about the measured ROP. Then Machine-learning models are used for its predictions to serve as a reference to evaluate any deviation and its possible causes, by testing the prediction in a new data set. This thesis is divided into four main parts. Starting by exploring data mining functionalities and its applications, including specific examples related to the Oil & Gas (O&G) industry. The following part involves understanding drilling data, its origins in measurements, its data type, and some of the challenges faced during its acquisition process. The ROP measurement is discussed in detail during this stage as well. With a general overview of the resources, the third part is dedicated to the methodology by developing a workflow including Pre-processing and Processing of the data using a commercial data mining software to implement a model for ROP prediction. In the last part, the Data Analysis and Model Evaluation are performed using different visualization tools, reinforced by descriptive statistics. A discussion of the model implementation and testing process is presented as well, based on the obtained results. The outcome of this work, drawn a road for further research on ROP deviation causes. It offers an insight for data mining applications for practical analysis and prediction derived from drilling data. It endorses its application when objectives are clearly defined and with no resources constraints.