A robotic manipulator, commonly employed for tasks such as object packing, sorting, and assembly, must be capable of handling items of varying shapes, sizes, weights, and colors. These tasks often require reprogramming especially when new manipulation challenges arise, making traditional approaches impractical. In this study, we propose the integration of two powerful decision-making and problem solving methodologies; Case-Based Reasoning (CBR) and Genetic Algorithm (GA) to enhance adaptability and efficiency of robotics manipulation for new manipulation tasks. CBR addresses new manipulation challenges by retrieving solutions from previously encountered cases that closely resemble the current problem. However, when the retrieved solution is suboptimal, conventional CBR methods typically require manual intervention to revise, adapt, and retain solutions, making this process slow and inefficient. To overcome these limitations, we propose the use of GA to iteratively refine and optimize retrieved solutions through GA operators such as selection, mutation, and crossover until convergence on an optimal or highly effective manipulation strategy is achieved. GA evaluates each candidate solution based on predefined fitness criteria, such as manipulation success rate, efficiency, and precision where most fitted solutions are retained in the case base. By integrating GA with CBR, the decision making process becomes significantly more efficient, as GA can automatically optimize, validate, and approve solutions without the need for human intervention, thereby accelerating new task execution. Furthermore, GAs provide a robust mechanism for handling complex and unforeseen manipulation tasks, enabling the system to adapt to highly dynamic tasks and environments and improve over time by automating case revision and retention process. This hybrid approach not only facilitates retrieval of past solutions but also continuously learns, optimizes, and retains new solutions, making the robotic system progressively smarter and more autonomous with each new manipulation challenge. Additionally, the proposed framework is scalable to increasingly complex tasks and environments, including scenarios involving large case bases and multiple collaborative robots operating within a shared workspace.