The virtual machine (VM) placement problem is a critical topic in cloud computing research, demanding collaborative decision-making through integrated consideration of multi-dimensional resources. To solve the problem that existing VM placement algorithms exhibit long turnaround time, load imbalance, and poor dynamic adaptability, this paper proposes a high-performance VM placement strategy, named DVMP. To enhance load balancing, the proposed DVMP strategy employs an adaptive weight adjustment mechanism combined with the TOPSIS algorithm to optimally select target hosts for each candidate VM. To adapt to dynamic environments, the strategy periodically collects and updates VM placement requests along with resource load status. To reduce turnaround time, the strategy employs multithreaded technology for concurrent VM deployment. The experimental results indicate that the placement efficiency of DVMP is improved by 41.5%, the cluster resource load imbalance is reduced by a maximum value of 67.1%, and the physical machine (PM) resource load imbalance is decreased by a maximum value of 39.0%, which can make the multi-dimensional resource utilization of the cluster more balanced.