Due to the great threat posed by excessive nitrite in food and drinking water to human health, it calls for developing reliable, convenient, and low-cost methods for nitrite detection. Herein, we string nanozyme catalysis and diazotization together and develop a ratiometric colorimetric approach for sensing nitrite in food. First, hollow MnFeO (a mixture of Mn and Fe oxides with different oxidation states) derived from a Mn-Fe Prussian blue analogue is explored as an oxidase mimic with high efficiency in catalyzing the colorless 3,3′,5,5′-tetramethylbenzidine (TMB) oxidation to blue TMBox, presenting a notable signal at 652 nm. Then, nitrite is able to trigger the diazotization of the product TMBox, not only decreasing the signal at 652 nm but also producing a new signal at 445 nm. Thus, the analyte-induced reverse changes of the two signals enable us to establish a ratiometric colorimetric assay for nitrite analysis. According to the above strategy, facile determination of nitrite in the range of 3.3–133.3 μM with good specificity was realized, providing a detection limit down to 0.2 μM. Compared with conventional single-signal analysis, our dual-signal ratiometric colorimetric mode was demonstrated to offer higher sensitivity, a lower detection limit, and better anti-interference ability against external detection environments. Practical applications of the approach in examining nitrite in food matrices were also verified.