To efficiently handle multimedia data transmissions across 5G mobile wireless networks, Quality-of-Service (QoS) techniques have been developed with statistical delay bounds. Recent developments in error-control technology have made Finite Blocklength Coding (FBC) an effective tool in the non-asymptotic regime assuring exacting statistical QoS standards in terms of both latency and reliability for the Ultra Reliable Low-Latency Communications (URLLC) in 5G networks, unlike Shannon's theorem which involves infinite blocklength. In order to address the aforementioned issues, under both asymptotic and non-asymptotic conditions, the paper proposes heterogeneous analytical-based resource allocation methods for 5G mobile networks predicated on Non-Orthogonal Multiple Access (NOMA) with massive Multiple-Input Multiple-Output (MIMO). It develops the NOMA with massive MIMO scheme-based 5G wireless mobile network model to maximize the performance and achieve interference management, high capacity, adaptive and intelligent optimization for the outlined strategies. It also executes simulation tests to demonstrate how well the suggested schemes perform in terms of ensuring heterogeneous statistical delay/error rate constrained QoS when compared to other current methods.