Human-driven environmental change has increased the occurrence of harmful cyanobacteria blooms in aquatic ecosystems. Concomitantly, exposure to microcystin (MC), a cyanobacterial toxin that can accumulate in animals, edible plants, and agricultural soils, has become a growing public health concern. For accurate estimation of health risks and timely monitoring, availability of reliable detection methods is imperative. Nonetheless, quantitative analysis of MCs in many types of biological and environmental samples has proven challenging because matrix interferences can hinder sample preparation and extraction procedures, leading to poor MC recovery. Herein, controlled experiments were conducted to enhance the use of ultra-performance liquid-chromatography tandem-mass spectrometry (UPLC-MS/MS) to recover MC-LR and MC-RR at a range of concentrations in seafood (fish), vegetables (lettuce), and environmental (soil) matrices. Although these experiments offer insight into detailed technical aspects of the MC homogenization and extraction process (i.e., sonication duration and centrifugation speed during homogenization; elution solvent to use during the final extraction), they centered on identifying the best (1) solvent system to use during homogenization (2-3 tested per matrix) and (2) single-phase extraction (SPE) column type (3 tested) to use for the final extraction. The best procedure consisted of the following, regardless of sample type: centrifugation speed = 4200 × g; elution volume = 8 mL; elution solvent = 80% methanol; and SPE column type = hydrophilic-lipophilic balance (HLB), with carbon also being satisfactory for fish. For sonication, 2 min, 5 min, and 10 min were optimal for fish, lettuce, and soil matrices, respectively. Using the recommended HLB column, the solvent systems that led to the highest recovery of MCs were methanol:water:butanol for fish, methanol:water for lettuce, and EDTA-NaPO for soils. Given that the recommended procedures resulted in average MC-LR and MC-RR recoveries that ranged 93 to 98%, their adoption for the preparation of samples with complex matrices before UPLC-MS/MS analysis is encouraged.