The importance of monitoring biodiversity to detect and understand changes throughout time and to inform management is increasingly recognized. Monitoring schemes should be globally unified, spatially integrated across scales, long term, and cost-efficient. We propose a framework to design optimized multispecies-targeted monitoring networks over large areas. The method builds upon previous developments on systematic conservation planning in terms of optimizing resource allocation in space, and comprises seven steps: (a) determine which questions will be addressed, (b) define species to be monitored, (c) compile occurrence data for all defined species, (d) predict the overall distribution of each species, (e) collect relevant environmental data and identify homogeneous strata, (f) set targets for the minimum number of monitoring sites per species and/or stratum and (g) identify optimal monitoring sites. We tested whether the monitoring networks designed with our framework have increased performance when compared to networks obtained with simple-random or stratified-random sampling by using a set of different indicators. To that end, we designed monitoring networks using optimized and non-optimized sampling schemes, applied to a case study in Portugal, where the goal was to design a monitoring network for amphibians and reptiles, to complement the one currently established in Spain. Results allowed us to conclude that monitoring networks designed with our method tend to outperform the non-optimized ones, in terms of higher species diversity (i.e. higher number of species and equity across monitoring sites), higher representation of environmental strata, and particularly higher coverage of rare species, with less survey effort. Synthesis and applications. We developed a framework to allocate monitoring sites for multiple species at broad scales using predictive models and optimization algorithms currently applied in systematic conservation planning. This framework presents field survey cost-efficiency advantages when compared to other standard sampling designs and can significantly contribute to improving the design of monitoring schemes. Thus, we recommend its application to design new multispecies monitoring networks or to extend existing ones.