<p>While the asking data rate demand grows higher, a new perspective with the usage of mmWave technology is proposed by the network system of fifth generation New Radio (5G NR) which offers low latency along with higher bandwidth and thus can fulfill the requirement the current generation cellular communication needs. However, utilizing mmWave frequencies involves more gNodeBs (5G base stations) since high-frequency signals cannot travel far, which means that handovers (HOs) occur more frequently; this presents a challenge to ensure smooth data flow for both stationary and mobile users. An increase in HOs increases the risk of ping-pong events, interruptions of data flow, and in the worst-case scenario, can result in complete radio link failure (RLF). The main purpose of this research is to increase handover efficiency by lowering the quantity of unnecessary and delayed handovers while sustaining high received signal quality. To achieve this objective, every handover-related parameter was examined; through MATLAB simulation, a feasible range of critical parameters was determined, such as handover margin (HOM) and time to trigger (TTT). Then all the related data was fed to a deep hybrid model which is built by combining long short-term memory (LSTM), gated recurrent unit (GRU), convolutional neural network (CNN), and dense neural network (DNN). The outcome from the hybrid model is promising as it could be able to reduce unnecessary HOs by up to 60%, and handover delay was reduced significantly while maintaining strong signal strength which was reflected in the simulation. As deep learning models heavily depend on the dataset, if the dataset is well obtained this type of approach has the potential to be an appropriate way to make handover efficient. </p>