Candidatus Liberibacter solanacearum (Lso) is the causal agent of zebra chip of potato (Solanum tuberosum), which can significantly reduce potato yield. In this study, a loop-mediated isothermal amplification (LAMP) method for the detection of Lso haplotypes A and B was developed and evaluated. Two sets of LAMP primers named LAMP-A and LAMP-B were designed and tested for specificity and sensitivity. Both LAMP-A and LAMP-B were specific to Lso in in silico analysis using the Primer-Blast tool. The LAMP-A and LAMP-B could only produce positive signal from DNA mixtures of Lso-infected tomato but not from the genomic DNA of 37 non-target plant pathogens. The sensitivity of LAMP-A and LAMP-B on Lso haplotypes A and B were tested on gBlocks and genomic DNA from Lso-infected tomato. On the genomic DNA, for LAMP-A, the lowest amount of template DNA for a positive LAMP reaction was 2 to 20 ng on four haplotype A strains and 20 to 80 ng on four haplotype B strains; for LAMP-B, the lowest amount of template DNA for a positive LAMP reaction was 0.02 to 2 ng on four haplotype B strains and 20 ng to no amplification on four haplotype A strains. On gBlocks, for LAMP-A, the lowest number of copies for a positive LAMP reaction was 60 on haplotype A and 600 on haplotype B; for LAMP-B, the lowest number of copies for a positive LAMP reaction was 60 on haplotype B and 600 on haplotype A. Therefore, considering the convenience of the LAMP technique, as well as the high specificity and sensitivity, the LAMP-A and LAMP-B primers can be used together to test the probable Lso-infected plant or psyllid samples to rapidly, accurately and directly differentiate haplotypes A and B. We highly recommend this LAMP system to plant pathology practitioners and diagnostic labs for routine detection of Lso and confirmation of zebra chip disease on potato or tomato.