All over the world, many offshore structures have been built in the last decades, both by the oil and gas industry and by the renewable industry. When these structures reach their end of life, they have to be removed. The removal is done in decommissioning projects. Many of those offshore structures are bottom founded. Of that amount, a big part has steel space frame substructures called jackets. Jackets are often removed with a heavy lift vessel. The crane hook of such a vessel is connected to the structure to be lifted by means of lift points. These lift points are the connection point of the rigging with the structure. The structures removed in decommissioning projects are often more than 30 years old. Because of that, there are a lot of uncertainties encountered by the industry in decommissioning projects that can lead to issues, such as loss of information or unknown material strength. The result is that all aspects of a decommissioning project have to be reconsidered each time. This makes the preparation phase costly and time-consuming. An aspect of the preparation is the selection of the lift point type, while it is unclear how and whether an optimum solution is chosen. Therefore the question arises, if the decision making process can be optimized by improving the selection of the lift point type in the preparation phase. One way to optimize such a decision-making process is by means of artificial intelligence. AI is able to compare data where humans see no connection, it can analyze large pieces of data in a much shorter time than humans can, it makes unbiased decisions and it can significantly reduce errors and increase accuracy and precision. This thesis will therefore focus on the optimization of the decision-making process of the lift point design of jackets by means of artificial intelligence. To answer the overall research question, first an analysis is made of various decommissioning projects of the past years. These projects were carried out by Boskalis. In this analysis, an attempt is made to visualize which steps are taken in the lift point selection process. By means of these steps it can be investigated whether there is a repetition or trend in the selection process of the lift points. In addition, the analysis aims to find the criteria the decisions are based on. Secondly, the analysis shows that certain steps are taken by Boskalis in the decision-making process of the lift point design in decommissioning projects. A trend was clearly visible in the way the lift points were selected. The relatively simple options are examined first, after which the more difficult options are investigated if the others prove impossible. In addition, the analysis found clear criteria that are used in the selection of the lift points. These criteria were reflected in the selection process of the execution strategy and in the selection process of the lift points. In addition to the total costs, among others ease of installation and the total duration also play a major role in the decision-making process. For example, an option was chosen that was more expensive than its alternative, but had a much shorter duration. This information made it possible to determine whether AI could be used in decommissioning projects in the future, to accelerate the preparation phase. Thirdly, it is investigated whether AI can improve the decision-making process of the lift point design. Therefore, three AI applications are examined and compared with the previous findings. The AI applications are: machine learning, genetic algorithms and single variable optimization. All three turned out not to be a good solution to the problem. For machine learning there is too little data available. Genetic algorithms do not seem to give the desired output with the given inputs. And single variable optimization is too limited in its solution. The final conclusion is therefore that the three investigated AI applications in this way will not improve the decision-making process of the lift point design. If the data set were to be expanded, machine learning could eventually be applied in the future. In addition, it could be investigated in future research whether genetic algorithms can be used if the input is changed. This thesis only briefly discusses what GA entails and whether the method has the desired output with the data found. Also, only a limited number of AI applications have been investigated in this thesis. But there are many today. For example, multi variable optimization could be a starting point for further research. Because there, the optimization function can have two variables, so that, for example, time and cost can be entered in different units.