Water is indispensable for life, yet many lack access to clean drinking water, resulting in fatalities from waterborne bacterial infections. Precise assessment of microbial abundance and viability in natural aquatic environments is vital. Adenosine triphosphate (ATP) serves as a parameter for viability assessments due to its presence in viable bacterial cells as an energy carrier. Traditional ATP detection methods involve chemical or enzymatic extraction, followed by measurement of light emission via the Luciferin–Luciferase complex. However, these methods are costly, present a low stability, require specialized equipment, and entail complex sample pretreatment. To overcome these limitations, we developed a biosensor based on aptamers, nucleic acid sequences with specific target-molecule-binding capabilities. Aptamers offer advantages such as an enhanced stability, a lower cost, and ease of design compared to antibodies. Recently, ATP has been used for aptamer selection testing. Our proposed biosensor utilizes a structure-switching ATP-binding DNA nanoswitch with two functional domains: a catalytic DNA-zyme domain and an ATP-binding aptamer domain. In the presence of ATP, its binding to the aptamer domain triggers the activation of the DNA-zyme domain, which is exploited for chemiluminescence (CL) detection. Integrating functional DNA biosensors with microfluidic paper-based analytical devices (µPADs) holds promise for point-of-care (POC) applications. However, achieving proper DNA binding on paper remains challenging, often requiring solution-based assay protocols, leaving µPADs for final signal readout. Here, we introduce an origami µPAD with preloaded dried reagents, allowing for on-paper assay execution upon sample addition and proper folding. Paper functionalization strategies and assay protocols were optimized to ensure simple and straightforward detection of ATP, employing a portable charge-coupled device (CCD) camera for CL detection. Calibration curves plotted against the logarithm of ATP concentration in the range of 1 to 500 µM facilitated determination of the assay's limit of detection (LOD), which was found to be 3 µM.