Multiagent reinforcement learning (MARL) primary-secondary control has implemented effective recent management of battery energy storage systems (BESSs), especially in direct current (DC) microgrids (MGs). However, there was an inaccurate charge/discharge synchronization and output voltage imbalance of the BESSs, especially under real operation challenges, such as high/continuous load fluctuations, uncertain BESSs’ insertions/removals (Plug-and-Play (PaP)), batteries’ heterogeneity/degradation, infrastructural/operational faults, infrastructure influence, and environmental impacts (temperature disparities). The defect is identified as a tradeoff between utilizing the real-time power storage capacity and charge/discharge synchronization accuracy, which results in circulating eddy currents, overloading, and imbalanced participation levels of the BESSs. The downsides are defective control stability, unsettled power flow balance, imperfect batteries’ health/life, and compromised introduction/sustainability of renewable energy of the MG. This research suggests, implements, and investigates a developed MARL-based control solution to improve BESS’s violated charge/discharge synchronization accuracy and output voltage imbalance, especially in critical operations and under the described real-operation influences. This is through the following contributions, which are preceded by a comprehensive and extensive literature review: 1) Enhance BESSs’ charge/discharge synchronization under real operation influences through developing a multiagent participation level correction approach (develop droop and local control and modify current consensus). 2) Improve output voltage balance through a new collaborative correction (develop secondary voltage control and modify voltage consensus correction). 3) Accomplish independent power management from the number of BESSs through new participation-based management of the BESSs’ insertions/removals (collaborative role of the multiagent and the developed participation correction approach). 4) Correct synchronization accuracy violations due to infrastructural/operational influences through a new real-time collaborative secondary correction and a multiagent state of charge (SOC) regulation. 5) Support BESS’s participation balance (enhance charge/discharge synchronization) under real operation influences through a developed multiagent compensation of the DC influence on the control due to infrastructural/operational influences. 6) Improve PaP performance and output voltage balance, reduce losses, enhance batteries’ health/life, and assess real-time power flow efficiency through a new economic power flow monitoring/organizing based on the achieved active/reliable PaP. The In-Lab online hardware-in-the-loop (HIL) interactions between the system model and the real-time platform (dspace-1202-2023X) confirm the outperformance of the proposed adaptive strategy, especially in critical operations. Enhanced PaP performance (reduced convergence time) by (18-62%), improved output voltage steadiness by (6.458-21.628%), better stability of battery’s voltage by (2.9739-3.8462%), developed load steadiness by (6.666-37.091%), reduced power consumption by (2.94%), enhanced power flow balance by (6.468%), and improved power flow efficiency by (2.626%), are the proposed strategy merit’s evidence.