As technology advances, the traffic demand has increased in tandem with the evolution of mobile servicerntypes, and prompting mobile operators to enhance their network capacity. Because optimum link capacityrnplanning is expressed in terms of both cost and network performance, operators face a challenge inrnmanaging limited network resources such as microwave channel frequency and modulation order. Althoughrnthe peak capacity of mobile services for provisioning of uplink microwave backhaul has resultedrnin underutilized links, it is a prevalent practice among service providers due to its ease of installation.rnUnderutilized microwave links required higher transmission power and antenna size due to excessivernmodulation order, and frequency channel spacing which decreases receiver sensitivity, that affects linkrnreliability. Thus, the fundamental purpose of this study is to optimize underutilized microwave link capacityrnusing a real-time traffic to free up unused resources. Hence, this paper proposed an approach ofrnlink capacity dimensioning using Probabilistic approach for busy hour traffic, and SARIMA time seriesrnmodeling for future traffic trend of sample sites from Addis Ababa. Therefore, the studyâ€™s key findingsrninclude 40% of sites in sample space have been reduced the unused channel frequency by half, 92% ofrnsites optimized their modulation order by one forth, and improvement of overall link percentage utilizationrnfrom 31% to 60% as minimum, as well as link reliability improvement of the microwave backhaulrnin particular, comparatively with the existing link capacity using the proposed approach.