Cluster analysis is a key technique in data exploration. With the rapid growth of modern network environments, the increasing diversity and volume of data attributes have reduced the effectiveness of traditional clustering methods. Segmentation-based clustering is preferred for its efficiency and simplicity compared to techniques like K-means. However, most studies focus only on numerical and categorical attributes or solely on sequential attributes, often leading to inaccurate clustering results or high computational costs. There is still room to enhance clustering efficiency and accuracy. This study incorporates three types of attributes and applies distinct distance measurement definitions. Categorical attributes are grouped based on common occurrence values, while sequential attributes are adjusted using a correction factor to address scale inconsistencies. The clustering performance is optimized by minimizing intra-cluster distances while maximizing inter-cluster differences. These three attribute types are treated as independent blocks, with group centers arranged and optimized accordingly. A genetic algorithm is introduced to determine the optimal combination of group centers. By integrating all three attribute types into a unified clustering model, this study improves clustering accuracy by separately processing sequential and numerical attributes. The use of genetic algorithms further enhances clustering efficiency, making the proposed approach more effective for analyzing complex datasets.