The reliability of wireless telecommunication service has become a major concern for the operation and maintenance (O&M) departments of the major telecommunication service providers. Consequently, reliability prediction has assumed a pivotal role for O&M department to furnish dependable services while simultaneously curtailing operational costs. However, predicting the service reliability is a challenging issue mainly attribute to the mobility characteristics of the end-users in the network. In tandem with the formulation of service reliability metrics, this study introduces a linear distribution regression model under Gaussian distribution and a kernel-based distribution regression model tailored for bimodal scenarios. The empirical validation of these proposed metrics and methodologies is substantiated through case studies employing both simulated data and real-world datasets derived from a wireless telecommunication system within an urban district. The outcomes of these case studies demonstrate the efficacy and applicability of the proposed metrics and methods.