Vehicular ad hoc network (VANET) becomes an important area of forthcoming 6G network. The utilization of 6G networks enables the achievement of maximum availability, reliability, and performance in the VANET. At the same time, accurate and timely traffic flow prediction is essential to accomplish enhanced quality of services in VANET. The latest developments of artificial intelligence (AI) models are found useful for the design of precise traffic flow forecasting approaches to assist drivers and travelers. This study develops a novel Henry gas solubility optimization with deep learning enabled traffic flow forecasting (HGSODL-TFF) technique for 6G enabled vehicular networks. The presented HGSODL-TFF technique primarily intends to forecast the level of traffic in the 6G enabled VANET. In addition, the HGSODL-TFF model initially preprocesses the traffic data using z -score normalization approach. Besides, deep belief network (DBN) model is employed to effectually forecast the traffic flow. Also, HSGO algorithm can be applied for optimally modifying the hyperparameters (such as learning rate, epoch count, and batch size) of the DBN model thereby improving the forecasting performance. The experimental validation of the HGSODL-TFF model is performed on test data, and the results are inspected under several aspects. The simulation results reported the betterment of the HGSODL-TFF model over the other recent approaches.