Ecological complex networks are common in the study of patched ecological systems where evolving populations interact within and among the patches. The loss of the dispersal connections between patches due to reasons such as erosion of migration corridors and road construction can cause an undesirable partitioning of such networks resulting in instability or negative impact on the metapopulations. A partitioning or spatial cut that is aware of the stability of the dynamics in the resulting daughter sub-networks can be an effective tool in dealing with the situation like proposing road alignment through a metapopulations network. This paper provides some mathematical conditions along with an heuristic graph partitioning algorithm that can help in finding ecologically suitable partitions of the metapopulations networks. Our study noted the crucial role of network connectivity (measured by Fiedler value) in stabilizing the metapopulations. That is, a sufficiently connected metapopulations network along with constrained internal patch dynamics has stable dynamics around its homogeneous co-existential equilibrium solution. With the considered mathematical model in this paper, network partitioning does not alter the internal patch dynamics around its homogeneous equilibrium point, but it can change the connectivity levels in the partitioned subnetworks. Thus, the proposed partitioning problem for an already stable metapopulations network is reduced to finding its subnetworks with desirable connectivity levels.