Visual electrochemiluminescence (ECL) biosensors do not need complex instruments or well-trained operators, which is regarded as an ideal choice for portable and low-cost detection. But traditional visual ECL biosensors are based on the change in ECL intensity, which is easily affected by environmental factors and signal acquisition processes. In this work, a visual ECL biosensor chip based on distance readout has been developed for the first time. The chip is composed of a square detection region and a visual channel region, which are modified with graphene oxide (GO) and gold nanoparticles (AuNPs)@Ti3C2 nanocomposites, respectively. Target molecules can release aptamers adsorbed on the GO surface of the detection region and further change the electrode potential of the visual channel region, which can determine the length of the long channel that generates visible ECL signals. The application of AuNPs@Ti3C2 nanocomposites can effectively enhance ECL intensity by six times. Through the unique design of the sensor chip, quantification detection can be achieved based on the length change instead of the traditional intensity change. This visual ECL sensor is successfully applied for deoxynivalenol toxin detection in actual samples, demonstrating that the introduction of the distance readout strategy into ECL sensing has a good prospect in on-site testing.