A significant proportion of vector-borne diseases are transmitted by blood-sucking dipterans, including mosquitoes. Understanding transmission risks requires accurate identification of species across heterogenous habitats, but many cryptic and polymorphic species are overlooked when using morphological identification. Estimates of mosquito diversity are typically based on adult female trapping methods which tend to target host-seeking species and may represent a biased snapshot of community structure. Unfortunately, diversity estimates based on larval data are rarely included in mosquito ecological analyses. We carried out adult and larval sampling over six months in Singapore using an integrative approach of morphological identification and molecular delineation with mini-barcodes (313 bp) generated on a Next Generation Sequencing platform to obtain species estimates. We collected 3201 mosquitoes across 58 species (14 genera). Notably, 16 species were collected only through larval sampling and 22 species were only resolved using mini-barcodes. Of the latter we identified three morphologically similar species groups and documented several intraspecific polymorphisms. We compared adult-only data against a full dataset (adult + larval + mini-barcode). The species accumulation curves reached an asymptote for all but one site when using the latter; non-metric multidimensional scaling (NMDS) revealed that mosquito communities were only well separated when using the full dataset. Overall, the latter reflects a more defined and accurate community structure across all sites. We find that several mosquito species were generally influenced by tree cover, rainfall and presence of large water bodies, further supporting the idea that many species are niche-specific. Synthesis and applications. We report the first successful use of mini-barcodes on mosquitoes and demonstrate its utility in delineating multiple challenging species groups. We recommend the use of both morphological and molecular identification methods for ecological studies and vector surveillance. Misidentification in species estimation, especially for medically relevant insect groups can lead to conflicting reports and slows down vector control efforts. We provide evidence that varying sampling techniques, particularly of the larval stages for holometabolous insects, is important in generating a robust dataset for downstream analyses. Together with DNA barcoding, this integrative approach helps to minimize error cascades when designing management strategies.