In resource-constrained network function virtualization (NFV) environments, it is expected that data throughput for service chains is maintained by using virtual network functions (VNFs) effectively. In this paper, we formulate an optimization problem for maximizing the total data throughput in resource-constrained NFV environments. Moreover, based on our formulated optimization problem, we propose a heuristic service chain construction algorithm for maximizing the total data throughput. This algorithm also determines the placement of VNFs, the amount of resources for each VNF, and the transmission route for each service chain. It is expected that the heuristic algorithm can construct service chains more quickly than the meta-heuristic algorithm. We evaluate the performance of the proposed methods with simulations, and we investigate the effectiveness of our proposed heuristic algorithm through a performance comparison. Numerical examples show that our proposed methods can construct service chains so as to maximize the total data throughput regardless of the number of service chains, the amount of traffic, and network topologies.