Tunable laser has been widely introduced in the field of optical fiber sensing as core component for spectrum analysis, wavelength division multiplexing and grating demodulation in recent years. There is a new type called Modulated Grating Y-branch (MG-Y) laser, and the characteristics of small size, wide tunable range and fast wavelength switching made it extremely popular in the field of optical fiber sensor demodulation. The primary key technology for MG-Y laser that needs to be solved in the application of fiber sensing is to control multiple input currents at the same time and build a stable, accurate and continuous wavelength Look Up Table (LUT). However, the wavelength tuning characteristics of MG-Y lasers are complex, the traditional methods are inefficient, the tuning wavelength can not meet the resolution requirements, and it is difficult to ensure the accuracy and stability of the output wavelength, which in turn affects the demodulation fiber sensing accuracy. Therefore, it is particularly necessary to find a new method for quasi-continuously tuned LUT construction.Aiming at the problem that it is difficult to achieve a stable quasi-continuous tuning of a specific wavelength for modulated grating Y-branch lasers, we proposed a method of constructing a wavelength look up table for MG-Y lasers. The method based on the KNN model, which is a simple classification algorithm. It can convert multi-class problems into multiple two-class problems for discussion, and calculate the accuracy rate, recall rate, and F¹-score. The construction process of LUT is mainly divided into 4 parts: obtain the initial scan data, identify the I_LR and I_RR quasi-continuous tuning range, identify the linear tuning range of each segment of I_PH, interpolation retrieval of target wavelength in a single path, generate LUT of MG-Y laser. This method has achieved rapid classification of the wavelength quasi-continuous tuning region. It used toolkit of sklearn machine learning. Its input is the wavelength parameter-control of L₁, output wavelength, which are converted into training set and test set according to 7∶3. And the data set of L₂ was used as verification set. Its output is the prediction category of each test sample. This method used euclidean distance to reflect the similarity of two example points, and the optimal K value is obtained by cross verification. At the same time, the relationship between F₁ value and K value is selected for experimental verification. The results show that when K=5, the classification result is better. Calculation time of the entire algorithm on the CPU is only 10 s, which has greatly reduced hours with an accuracy rate of 77%. According to the model prediction result, its maximum wavelength interval obtained is 50 pm. This method can meet the needs of engineering applications and generated a smooth, which made the best of phase tuning characteristics of MG-Y laser and adopted Newton interpolation to fine-tune the wavelength.To verify the application value of the method proposed in this paper, the following experiments are carried out in turn. Firstly, to verify the accuracy and stability of MG-Y laser, we built a laser LUT experimental system, and continuously scanned the LUT which based on our method 20 times. According to LUT, the laser drive current is controlled to output a specific wavelength of light. The experimental system used Xilinx XC7Z020 as the main control unit, and used a multi-wavelength meter (AQ6151, with an accuracy of 0.3 pm, its sampling frequency is 1 Hz) to acquisition output wavelength. The experimental results show that its accuracy is about 2 pm, and the stability is 0.7 pm. Secondly, to verify the effectiveness of the constructed LUT, we built the F-P etalon wavelength demodulation experiment system. We placed a F-P etalon in the high and low temperature test chamber (GDW-100), the temperature was kept at 25 ℃. The sampling frequency of demodulation system is 250 Hz. The experimental results show that the F-P etalon demodulation wavelength stability is 1.73 pm. Finally, to verify the effectiveness of LUT in practical engineering applications, we built an FBG wavelength demodulation experimental system, which based on the detection requirements of the three-way load of a certain type of aircraft landing gear. Under impact conditions, the FBG sensor network was used to obtain the strain response of the landing gear structure, and the MG-Y laser-based demodulation system was used to collect and demodulate the FBG strain spectrum information. Its single sampling time is 5 s, and its sampling frequency is 250 Hz. By analyzing the data collected for 6 consecutive days, we come to the conclusion that the FBG demodulation wavelength stability is 1.75 pm, and the FBG demodulation wavelength correlation coefficient R is greater than 0.952 5. The experimental results proved that this method can control the MG-Y laser to achieve stable wavelength quasi-continuous tuning. In summary, the fiber grating demodulation system, which used this method, can realize stable F-P etalon and FBG spectrum acquisition and wavelength demodulation. This method has good practical application value.