Natural fractures and pores in carbonate reservoirs are highly developed, but their connectivity is poor and heterogeneity is strong. Traditional acid fracturing has a limited number of communication reservoirs, a low probability of communication, and a simple and single fracture morphology. Therefore, it is necessary to transform the traditional one-way communication mode into a targeted communication mode, increase the probability of communication, achieve all-round communication around the well storage collective, and achieve complex artificial fracture network structures. However, the mechanical mechanism of the initiation and propagation of volume fractures in fractured carbonate rocks is still unclear during the fracturing process. In order to form a scientific understanding of the expansion morphology and mechanism of volume fractures in fractured carbonate rocks, an analytical model for the propagation of complex fractures in fractured reservoirs was established. Physical simulation experiments on the initiation and propagation of volume fractures in fractured carbonate rocks were conducted, and a neural network algorithm for complex fracture networks was constructed. The optimal construction diagram for multi factor coupled connected fractures and grooves was drawn to guide the optimization of volume fracturing technology and improve fracturing effectiveness.