Electrophoresis titration (ET) based on the moving reaction boundary (MRB) theory can detect the analyte contents in different samples by converting content signals into distance signals. However, this technique is only suitable for on-site qualitative testing, and accurate quantification relies on complex optical equipment and computers. Hence, applying this method to real-time point-of-care testing (POCT) is challenging. In this study, we developed a smartphone-based ET system based on a visual technique to achieve real-time quantitative detection. First, we developed a portable quantitative ET device that can connect to a smartphone; this device consisted of five components, namely, an ET chip, a power module, a microcontroller, a liquid crystal display screen, and a Bluetooth module. The device measured 10 cm×15 cm×2.5 cm, weighed 300 g, and was easy to hold. Thus, it is suitable for on-site testing with a run time of only 2-4 min. An assistant mobile software program was also developed to control the device and perform ET. The colored electrophoresis boundary can be captured using the smartphone camera, and quantitative detection results can be obtained in real time. Second, we proposed a quantitative algorithm based on ET channels. The software was used to recognize the boundary migration distance of three channels, a standard curve based on two given contents of the standards was established using the two-point method, and the content of the test sample was calculated. Human serum albumin (HSA) and uric acid (UA) were used as a model protein and biosample, respectively, to test the performance of the detection system. For HSA detection, different HSA solutions were mixed with a polyacrylamide gel (PAG) stock solution, phenolphthalein was added as an indicator, and sodium persulfate and tetramethyl ethylenediamine (TEMED) were used to promote polymerization to form a gel. For UA detection, agarose gel was filled into the ET channel, the UA sample, urate oxidase, and leucomalachite green were added into the anode cell and incubated for 20 min. ET was then performed. The fitting goodness (R2) values of HSA and UA were 0.9959 and 0.9935, respectively, with a linear range of 0.5-35.0 g/L and a log-linear range of 100-4000 μmol/L. The limits of detection for HSA and UA were 0.05 g/L and 50 μmol/L, respectively, and the corresponding relative standard deviations (RSDs) were not greater than 2.87% and 3.21%, respectively. These results demonstrate that the detection system has good accuracy and sensitivity. Clinical samples collected from healthy volunteers were used as target blood samples, and the developed system was used to measure serum total protein and UA levels. Serum samples from five volunteers were selected, standard curves of total serum protein and UA were established, and the test results were compared with hospital standard testing results. The relative errors for serum total protein and UA were less than 6.03% and 6.21%, respectively, and the corresponding RSDs were less than 3.72% and 5.84%, respectively. These findings verify the accuracy and reliability of the proposed detection system. The smartphone-based ET detection system introduced in this paper presents several advantages. First, it enables the portable real-time detection of total serum protein and UA. Second, compared with traditional ET strategies based on colored boundaries, it does not rely on optical detection equipment or computers to obtain quantitative detection results; as such, it can reduce the complexity of the operation and provide portability and real-time metrics. Third, the detection of two biomarkers, serum total protein and UA, is achieved on the same device, thereby improving the multitarget detection potential of the ET method. These advantages render the developed method a promising detection platform for clinical applications and real-time POCT.