This study focused on the detection and quantification of eight cyanotoxins in water samples in three reservoirs located in the eastern department of Antioquia, Colombia. The reservoirs are a source of water supply and hydroelectricity, and also generate economic activities in fishing and recreation. Between May 2015 and October 2016, 8 samplings were carried out at times of high temperatures, which ranged from 20 to 29 °C. This period was selected because of a significant or strong El Niño phenomenon, according to the World Meteorological Organization. For the study, 270 integrated samples were taken from the photic zone (PZ) and the surface of the reservoirs, at each of the three sampling points. The samples were analyzed by the analytical technique of ultra-high-performance liquid chromatography coupled to triple quadrupole mass spectrometry (UHPLC MS/MS). The quantification performed for six microcystins (MCs), a nodularin (NOD) and a cylindrospermopsin (CYN), showed positive results well above 1 μg L−1. In the water of the Abreo Malpaso and Peñol reservoirs, microcystin-LR (MC-LR), microcystin-YR (MC-YR) and [D-Asp3,(E)-Dhb7]- microcystin-RR toxins were detected at levels of considerable concentration, especially between May and September 2015, when there was no rainfall in this region. In the Playas reservoir, positive results for [D-Asp3,(E)-Dhb7]-MC-RR were detected from May to November 2015, with the highest concentrations being reached in dry season. The temperatures reached and the changes in climatic conditions witnessed during the monitoring period of this study were important factors in the production of cyanotoxins. This was evidenced in this work by the high concentrations of detected cyanotoxins and their absence in periods of rain, as happened from the second quarter of 2016 until the end of the study in October. This is the first study of the detection and quantification of cyanotoxins in tropical reservoirs of northeastern Colombia using the UHPLC MS/MS analytical technique, which allowed the toxins to be unequivocally detected and confirmed. A method was developed and validated, proving to be sensitive, reproducible and accurate. For each of the toxins (microcystin-LR (MC-LR), microcystin-RR (MC-RR), microcystin-YR (MC-YR), [D-Asp3,(E)-Dhb7]- microcystin-RR, microcystin-LW (MC-LW), microcystin-LF (MC-LF), nodularin (NOD) and cylindrospermopsina (CYN)) the correlation coefficients (R2) were in a range between 0.9907 and 0.9999. Verification of the accuracy of the method was performed through a calibration curve in solvent. The recovery percentages of the accuracy and precision tests of the method for low level, medium level and high level were in a range between 64% and 115% for all the cyanotoxins. The validation of the cyanotoxin method shows that it is possible to detect them individually in natural water with a quantification limit (LOQ) of approximately 0.05 μg L−1.