This study presents a method for the prediction of equivalence ratio in pulse combustor from the ion current amplitude spectrum measured by a multi-frequency ion current sensor supplemented with principal component analysis (PCA) and support vector machine (SVM) technologies. An experimental study was conducted on a Helmholtz-type combustor to validate the proposed method. During the experiment, a multi-frequency excitation source whose crest factor was pre-optimized was applied on the ion current sensor to obtain ion current signals with high SNR. The excitation source has ten frequency components from 500 Hz to 5 kHz. The multi-frequency ion current signals were acquired under six different operation conditions. Under each condition, three typical positions with different distances from the gas inlet of the combustor were selected. The amplitudes of all frequency components in the multi-frequency ion current signals were extracted and projected into two subspaces via PCA to reduce the data dimensions and attenuate the noise corruption. An SVM model was built based on the amplitude spectrum features in the first two principle components subspace to predict the equivalence ratio. Experimental results show that the ion current amplitude spectrum relies on the equivalence ratio and the position where the sensor is installed. The equivalence ratio can be predicted with the PCA-SVM model with an average determination coefficient higher than 98%. This study validates the feasibility of the amplitude spectrum obtained by the ion current sensor for prediction of the equivalence ratio in the pulse combustor.